GTA Online might be one of the most popular and longest-lasting games to date, but it isn’t without its flaws. More specifically, GTA Online has an insane loading time of up to six minutes on most PC hardware (possibly more on slower PCs), which is absurd given how old the game is. A fix might be in the works, or at least, a fan of GTAV has figured out the problem and fixed it with a simple solution. Well, ‘fixed’ might be too strong a word, but the situation has improved at least.
Before all this, the GTAV fan “T0ST” was super annoyed at the absurdly long wait times he was getting in GTA Online. His load times were a whopping six minutes just to get into the game. To start troubleshooting, he started running Task Manager to see what was taking up so many resources during that 6-minute time frame.
He found out that the game was only loading two CPU cores consistently, for the entire duration of the load time. Nothing else was being utilized, so storage, graphics, and network bandwidth didn’t appear to be the bottleneck. This is unusual behavior, as a game should be loading stuff from storage into system memory and talking with the server, which could have at least partially explained the long wait times.
So T0ST delved deeper, into the game code itself, and eventually found the problem. It turns out the game is basically doing a ton of extra work for no reason at all. The game is basically looking at 10MB JSON file with a bunch of in-game store items, repeatedly, as it works to load the game. This is something the game can completely skip and do later. It also repeats scanning of the 10MB JSON every few bytes, further exacerbating the problem.
T0ST fixed this by creating his own DLL with optimized code and installing it into the game files. That cut load times down to a far more pallatable 1 minute and 50 seconds. That’s still pretty awful for a game that came out so long ago, but it’s one-third the time as before. Hopefully, this will encourage Rockstar to finally take a look this portion of the game code, and using data from T0ST it should be able to implement the fix in a future patch.
Samsung just leveled the playing field a bit in the battle between TVs and gaming monitors. PC monitors have long held an advantage over TVs for gaming, largely due to speedier refresh rates and response times. The best gaming monitors also fight screen tearing with some flavor of Adaptive-Sync. TVs, meanwhile, have made advanced display technologies, like OLED and mini LED, more attainable. Today, Samsung threw bones toward both corners, announcing the first TV with AMD’s most advanced screen tearing fighting technology and the marriage of its “Quantum Mini LED” technology and Samsung Odyssey G9 curved gaming monitor.
During its Unbox and Discover event today, Samsung debuted the first TVs to include AMD FreeSync Premium Pro, which will work with both PCs and gaming consoles, like the new PlayStation 5 and Xbox Series X. FreeSync Premium Pro will be available on Samsung’s 4K resolution Q70A and up, as well as the 2021 version of its customizable TV, aptly named The Frame.
There are numerous TVs available from Samsung, as well as from LG, with FreeSync Premium, which calls for a minimum 120 Hz refresh rate and adds low framerate compensation (LFC) compared to standard FreeSync. FreeSync Premium Pro takes things a step further by also supporting HDR content.
The vendor’s also throwing in a new Game Bar, which helps gamers monitor “critical aspects of play,” according to today’s announcement, and use Samsung’s Super Ultrawide Gameview feature, enabling ultrawide aspect ratios more commonly found in PC monitors than TVs.
But the electronics giant also had news to share with gamers committed to PC monitors, however. The Samsung Odyssey G9 needed no help being extreme. With a 1000R curve, it’s already as curvy as gaming monitors get today. However, Samsung upped the premium ante today by unveiling the 2021 version of the monitor with Quantum Mini LEDs. The 2020 version uses QLED, which is just another type of LEDs invented by Samsung for improved brightness and color.
Quantum Mini LEDs are already available in Samsung TVs, but the 2021 Odyssey G9 will mark one of the first mini LED gaming monitors, if not the first (depending on when it’s actually available to buy.).
According to Samsung, its Quantum Mini LEDs are 1/40th the height of a standard LED (what you’ll find in the vast majority of gaming monitors). Additionally, “instead of using a lens to disperse light and a package to fix the LED in place, [a] Quantum Mini LED has incredibly thin microlayers filled with many more LEDs.” Samsung’s Quantum Matrix Technology is supposed to enable precise control over those tiny LEDs to help fight blooming, also known as the halo effect. This would be particularly impressive, as even premium gaming monitors with FALD backlights can fall victim to the halo effect.
We already saw what the power of mini LEDs could do for image quality, particularly contrast, in our review of the Asus ProArt PA32UCX, a monitor for professionals. But we’ve yet to see the technology in a PC monitor built for gaming. Asus and Acer have both promised to release their own mini LED gaming monitors, but that hasn’t happened yet.
Intel might have finally set an official release date for its Rocket Lake chips that will soon vie for a spot on our list of Best CPUs. HardwareLUXX editor Andreas Schilling today reported that “the sales embargo for 11th Gen Desktop processors (Rocket Lake S) is March 30, 2021, at 6 a.m. PT / 9 a.m. ET.”
The announcement was largely symbolic. Representatives from MSI and Gigabyte revealed the March launch window in January, Milwaukee PC leaked the line’s pricing a few days ago, and Mindfactory sold its entire Core i7-11700K stock last weekend.
Here are the leaked prices courtesy of Milwaukee PC, but be aware that this might not reflect official pricing:
Processor
Pricing (Excl. VAT)
Cores / Threads
Base Clock (GHz)
L3 Cache (MB)
Part Number
Core i9-11900K
$599.99
8/16
3.5
16
BX8070811900K
Core i9-11900KF
$579.99
8/16
3.5
16
BX8070811900KF
Core i9-11900
$509.99
8/16
2.5
16
BX8070811900
Core i9-11900F
$479.99
8/16
2.5
16
BX8070811900F
Core i7-11700K
$484.99
8/16
3.6
16
BX8070811700K
Core i7-11700KF
$454.99
8/16
3.6
16
BX8070811700KF
Core i7-11700
$389.99
8/16
2.5
16
BX8070811700
Core i7-11700F
$359.99
8/16
2.5
16
BX8070811700F
Core i5-11600K
$309.99
6 / 12
3.9
12
BX8070811600K
Core i5-11600KF
$279.99
6 / 12
3.9
12
BX8070811600KF
Core i5-11600
$264.99
6 / 12
2.8
12
BX8070811600
Core i5-11500
$234.99
6 / 12
2.7
12
BX8070811500
Core i5-11400
$214.99
6 / 12
2.6
12
BX8070811400
Core i5-11400F
$179.99
6 / 12
2.6
12
BX8070811400F
Rocket Lake processors have popped up in benchmarks over the last few months, too, and some outlets recently published reviews of Core i7-11700K and Core i9-11900K samples to offer some idea of how Intel’s latest CPUs will perform.
It’s clear that the chips will place well on our CPU benchmarks hierarchy. Still, those reviews shouldn’t be taken as gospel—retail versions of the CPUs should offer greater performance than engineering samples—but they can help set expectations for Rocket Lake’s potential. (And its ability to compete with AMD’s offerings.)
Still, it’s good to hear that Rocket Lake should arrive this month, even if some retailers let the cat out of the bag a little early. We’ve reached out to Intel to confirm the March 30 launch plans and will update this post with the company’s response.
More information about all things Rocket Lake can be found right here.
We have with us the PNY GeForce RTX 3070 XLR8 Revel Epic-X RGB graphics card. The XLR8 (reads “accelerate”) line of graphics cards by PNY targets the gaming and PC enthusiast community and comes with the trust creative professionals associate with the PNY brand. PNY is a major vendor of professional visualization graphics cards based on NVIDIA Quadro graphics processors, as well as a leading vendor of flash memory products used by photographers and cinematographers. Palit-Gainward has been PNY’s manufacturing partner for these XLR8 graphics cards, and the RTX 3070 XLR8 Revel bears a striking resemblance to the Gainward RTX 3070 Phoenix GS.
The GeForce RTX 3070 is a fascinating performance-segment product that’s dancing on the edge of the enthusiast segment. NVIDIA claims it’s designed to perform better than the previous-generation flagship, the RTX 2080 Ti, but at half the price. This would mean the RTX 3070 has the same performance outlook as the RTX 2080 Ti—to enable maxed out gaming with RTX-on at 1440p resolution, 4K UHD gaming with RTX-off, or DLSS Quality enabled. It should also let e-sports gamers play at 1440p with high refresh rates.
At its heart, the RTX 3070 is based on the 8 nm “GA104” silicon, which it nearly maxes out. This is NVIDIA’s second-largest GeForce “Ampere” silicon. The company used a conventional memory setup by giving it 8 GB of 14 Gbps GDDR6 memory across a 256-bit wide memory bus, same as the RTX 2070. For the full details on the RTX 3070 technology and architecture, refer to our RTX 3070 Founders Edition article.
The GeForce Ampere graphics architecture heralds the 2nd generation of NVIDIA’s path-breaking RTX technology, which leverages real-time raytracing and AI deep-learning to dramatically uplift eye candy. It combines new Ampere CUDA cores capable of concurrent INT32+FP32 operations, 2nd gen RT cores that double the ray intersection performance over the 1st gen and add more hardware to accelerate newer kinds of raytraced effects, and 3rd gen Tensor cores that leverage the sparsity phenomenon to build and train AI deep-learning neural nets faster.
The PNY RTX 3070 XLR8 Revel Epic-X features a massive triple-slot product design with a triple-fan heatsink. The cooler is longer than the PCB underneath, so airflow from one of the fans flows right through because of vents on the backplate. The card also features a rugged industrial design, and a tastefully executed RGB LED illumination. Interestingly, our card sticks with NVIDIA-reference clock speeds, with its GPU Boost set to 1725 MHz and memory left untouched at 14 Gbps.
NASA’s brand-new Perseverance rover is the most advanced machine that’s ever landed on Mars. But when it comes to rovers, “state of the art” is a subjective term. Perseverance is running on none other than a PowerPC 750, a single-core, 233MHz processor with just 6 million transistors that’s most famous for powering the original “Bondi blue” iMac from 1998. It’s the same type of processor that NASA already uses in its Curiosity rover.
That may seem like a waste to some. After all, even with the difficulty of buying computer parts these days, surely NASAcould have found the budget for something like Intel’s $500 Core i9-10900K CPU (with 10 cores and a max clock speed of 5.3GHz) somewhere in the $2.7 billion cost of Perseverance. But as New Scientist explains, such an advanced chip is actually a detriment to the unique operating conditions of Mars.
That’s largely because Mars’ atmosphere offers far less protection from harmful radiation and charged particles than Earth’s atmosphere. A bad burst of radiation can badly wreck the sensitive electronics of a modern processor — and the more complex the chip, the more can go wrong. Plus, at 138 million miles away, it’s not like NASA can just swap out the processor if things go sideways. Because of those conditions, Perseverance actually features two computing modules: one is a backup just in case something goes wrong. (A third copy of the module is also on board for image analysis.)
To make the system even more durable, the PowerPC 750 chip in Perseverance is a little different than the one in the old iMacs. It’s technically a RAD750 chip, a special variant that’s hardened against radiation and costs upwards of $200,000. The chip is popular for spacecraft, too: in addition to Perseverance and Curiosity, it also powers the Fermi Space Telescope, the Lunar Reconnaissance Orbiter, the Deep Impact comet-hunting spacecraft, and the Kepler telescope, among others.
While the processor may be weak compared to a modern smartphone or gaming PC, NASA’s spec sheet for Perseverance notes that it’s far more powerful than earlier rovers like Spirit or Opportunity: its 200MHz clock speed is 10 times faster than those older rovers, and with 2GB of flash memory, it offers eight times the storage. (Rounding things out, Perseverance also has 256MB of RAM in case you were looking to build your own rover.)
But while the chip itself has been to Mars before, Perseverance features some new computer technology that’s debuting on the planet for the first time: Linux, which powers the Ingenuity helicopter that will attempt to fly autonomously on Mars as part of Perseverance’s mission.
Last week, Microsoft’s Alex Kipman, the inventor of Kinect and HoloLens, appeared in my living room to hand me jellyfish and sharks. That might sound like I had a weird dream, but it was a meeting made possible through Microsoft’s new Mesh platform. I donned a HoloLens 2 headset, joined a virtual meeting room, and Kipman immediately appeared next to my coffee table, ready to demonstrate Microsoft’s vision for the future of VR and AR — or as Microsoft calls it, mixed reality.
It all felt like a Microsoft Teams meeting set in the future.
Mesh is a collaborative platform that allows anyone to have shared virtual experiences on a variety of devices. “This has been the dream for mixed reality, the idea from the very beginning,” explains Kipman. “You can actually feel like you’re in the same place with someone sharing content or you can teleport from different mixed reality devices and be present with people even when you’re not physically together.”
Initially, Mesh will present people as virtual avatars taken from the AltspaceVR social network that Microsoft acquired back in 2017. Mesh will eventually support what Microsoft calls “holoportation,” allowing people to appear as themselves in a virtual space.
During my hour-long meeting in Microsoft Mesh, I constantly felt like this could be a far-future version of Microsoft Teams. Kipman appeared next to me as an avatar and started handing me virtual jellyfish and sharks. I could reshape the animals, pass them back, or just place them down in front of me. Although we weren’t working on some grand design or 3D model, it felt far more immersive than the Zoom video calls I have to attend on a near-daily basis.
It was the next best thing to having Kipman in the room with me and reminded me of my first experience with HoloLens. Microsoft originally demonstrated HoloLens using a collaborative Skype call where a technician could guide you on how to fix some wiring. It felt like the promise of augmented reality, and Microsoft Mesh seems like the natural next step.
“You can completely imagine a Mesh-enabled Microsoft Teams, where the key thing there is, think about colleagues from across the globe collaborating as if you and I are in the same physical location,” says Kipman. “Mesh enables Teams to allow organizations to essentially do mixed reality gatherings with everyone in the same room and so you should think about that in a Mesh-enabled Teams-type of an environment.”
Mesh isn’t just an app for holding virtual meetings, though; it’s an entire platform built on top of Azure that Microsoft hopes developers will tap into. Microsoft is hoping architects, engineers, and designers will all see the promise of Mesh, particularly during a pandemic when it’s difficult to work with 3D physical models without all being in the same room.
Microsoft is also making Mesh available on a variety of devices, including the HoloLens 2, most virtual reality headsets, tablets, smartphones, and PCs. A preview of the Microsoft Mesh app for HoloLens 2 will be available today, alongside a preview version of AltspaceVR that is Mesh enabled. Microsoft is planning to integrate Mesh into Teams and Dynamics 365 in the future, which might help bring the unique meeting experience I had into a reality for more people.
Microsoft certainly thinks Mesh will be the next big thing for mixed reality. CEO Satya Nadella compared the new platform to Xbox Live during his keynote speech at Ignite today. “Think about what Xbox Live did for gaming — we went from single player to multiplayer, creating communities that helped people connect and achieve together,” said Nadella. “Now just imagine if the same thing happened with mixed reality.”
The big thing holding Microsoft Mesh back is the cost of HoloLens devices and VR headsets. While you can jump in from a 2D screen like a phone or PC, it’s nowhere near as immersive as using a headset. Microsoft has tried to bridge this gap in the past with ambitious projects like Minecraft Earth, offering AR experiences on mobile phones. That particular project didn’t work out, and it’s not clear whether Microsoft Mesh will strike a chord with users either.
Microsoft Mesh will need some strong developer support or offer experiences you just can’t get anywhere else in VR to pick up momentum. Microsoft has recruited filmmaker James Cameron, Pokémon Go developer Niantic, and the co-founder of Cirque du Soleil to demonstrate the promise of Mesh during its Ignite keynote today.
A proof-of-concept version of Pokémon Go running on the HoloLens 2 and virtual concerts certainly show what’s now possible with Microsoft Mesh. The reality will now depend on the popularity of headsets or Microsoft’s ability to market this new technology toward businesses looking to embrace hybrid ways of working as the world navigates its way out of a pandemic.
Update, February 2nd 11:40AM ET: Article updated with comments from Microsoft CEO Satya Nadella.
At $329, Apple’s eighth-generation iPad is the most affordable tablet offered by the tech giant. But if you want to save some additional cash, you can buy the base model, which includes 32GB of storage, for only $299 at Amazon in space gray only (in stock March 28th). Walmart is price-matching the e-commerce giant, but you have the option to buy it for $299 in space gray or rose gold.
This iPad includes a 10.2-inch display and features an A12 Bionic processor. The tablet also includes a Touch ID fingerprint sensor and a Smart Connector, allowing you to connect it to a Smart Keyboard.
March is a great month for PS4 and PS5 owners. If you have an active PlayStation Plus subscription, you can snag a free digital copy of Final Fantasy VII Remake. If you need to buy a subscription or renew your membership this month, you can grab a one-year membership to the service for only $28 at Eneba.
Amazon and Best Buy have discounted Crucial’s P5 internal SSD. You can grab a 1TB configuration for only $110, the lowest price yet. It’s a good option for PC owners looking to upgrade their systems or improve their games’ loading times. Crucial says the P5 SSD promises up to 3400 MB/s read speeds and 3000 MB/s write speeds.
Crucial P5 1TB SSD
$110
$150
27% off
Prices taken at time of publishing.
A good SSD option for people looking to improve loading speeds on their PC games with up to 3400 MB/s read speeds and 3000 MB/s write speeds.
The range of Raspberry Pi Pico accessories keeps growing, and the latest is from Olimex. In a render teased via Twitter, Olimex appears to be working on their own breakout board with a focus on emulation, and it looks like they have crammed in everything we could need for a dose of retro computing.
RP2040-PICO-PC teaser small PC with RP2040-PICO module. Video, Audio, SD-card, UEXT, I2C, Lipo battery, Reset, USB pwr, USB host for keyboard, Debug TxRx, SWD for JTAG debug #rp2040 #raspberrypi #circuitpython #retrogames pic.twitter.com/str79xsMkmMarch 1, 2021
See more
Olimex, a well-known name in the maker community, has teased its own breakout board for the Raspberry Pi Pico. Its focus is on designing a board to accommodate retro emulation. Spurred on by Graham Sanderson’s BBC Micro computer emulator, Olimex confirmed that this board would be their first board in the Pico range.
The RP2040-PICO-PC is, as the name cryptically suggests, a ‘PC’ that uses the Raspberry Pi Pico as the heart of the board. Looking at the teaser image, it appears that Olimex is looking to use Pico boards directly with the RP2040-PICO-PC. But in a recent tweet, we saw that Olimex had received a sample of bare RP2040 chips intended for community members to develop their own RP2040 devices. Could we see an RP2040 at the heart of the board?
Back to the teaser. In the image, we can see series of two male headers for the Raspberry Pi Pico. Whether they are intended to hold the Pico via solder or friction fit remains to be seen. There is a chance that they could also be breakouts for the GPIO, and the RP2040 is mounted on the underside of the board. Right now, the render is all we have to go on.
The RP2040-PICO-PC has a micro USB port on the right side of the board, and next to there is a two-pin JST connector, most likely for a Lipo battery connection. There’s also a micro SD card connector on the bottom left of the board, most likely accessible via an SPI interface. Along the top edge of the board, we see a DVI video output, a USB host connector for a keyboard, and a 3.5mm audio jack. A 10-pin connector is right at the top of the board — this is a UEXT (Universal EXTension) connector that breaks out I2C, SPI, and RS232 serial. The UEXT connector is intended for use with UEXT modules, such as an LCD screen and sensors that are available separately.
How much and when can we buy one? Well, for now, those questions will have to remain unanswered, but we can say that the RP2040-PICO-PC could be a fun project for retro enthusiasts.
Choosing the best thermal paste may sound simple, but a dizzying array of options can turn this seemingly easy choice into a time-consuming task. Thermal paste is designed to minimize microscopic air gaps and irregularities between the surface of the cooler and the CPU’s IHS (integrated heat spreader), the piece of metal which is built into the top of the processor.
Good thermal paste can have a profound impact on your performance, because it will allow your processor to transfer more of its waste heat to your cooler and continue operating at a high clock speed for longer while a mediocre product will force your system to throttle in order to avoid overheating. Even the best CPUs are made better by good thermal paste.
Performance enthusiasts often swear by a favorite brand of paste, but new recipes crop up from time to time. We’ve conducted extensive testing of some of the top thermal pastes on the market, and if you’re on the hunt for more exotic pastes that are off the beaten path, we also have plenty of historical data below covering 90 different pastes (and yes, we’ve even tested toothpaste).
Thermal paste can go by many names, such as “thermal compound,” “thermal grease,” or “TIM” (Thermal Interface Material). Most pastes are comprised of ceramic or metallic materials suspended within a proprietary binder which allows for easy application and spread as well as simple cleanup.
These thermal pastes can be electrically conductive or non-conductive, depending on their specific formula. Electrically conductive thermal pastes can carry current between two points, meaning that if the paste squeezes out onto other components, it can cause damage to motherboards and CPUs when you switch on the power. A single drop out of place can lead to a dead PC, so extra care is imperative.
Liquid metal compounds are almost always electrically conductive, so while these compounds provide better performance than their paste counterparts, they require more focus and attention during application. They are very hard to remove if you get some in the wrong place, which would fry your system.
In contrast, traditional thermal paste compounds are relatively simple for every experience level. Most, but not all, traditional pastes are electrically non-conductive. We have a detailed breakdown of the best methods for applying thermal paste here.
Here are the best thermal pastes, along with our complete test results. Keep in mind that you don’t always get what you pay for as some budget pastes come awfully close to or overtake their more-expensive competitors.
Best Premium (Standard) Thermal Paste
1. ProlimaTech PK-3 Nano Aluminum
Best Premium
Electrically Conductive: No | Thermal Conductivity: 11.2 W/mk | Ease of Use: 4.5 (1=difficult, 5=simple) | Relative Performance: 4.0 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $3.60
Stability
Easy to use
Easy to clean up
None
Often hanging close to, but not overtaking, liquid metal compounds in our tests, PromilaTech PK-3 Nano Aluminum paste is a moderately viscous thermal compound integrated with aluminum and zinc oxide. The paste provides good stability, making it easy to apply to all surfaces as well as simple to clean up. You really can’t go wrong with this paste — it is sufficient for most CPU applications.
2. Thermal Grizzly Kryonaut
Runner-up: Best Premium Paste
Electrically Conductive: No | Thermal Conductivity: 12.5 W/mk | Ease of Use: 4.5 (1=difficult, 5=simple) | Relative Performance: 4.0 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $9.99
Stability
Easy to use
Easy to clean up
Price per gram
Thermal Grizzly Kryonaut is a mildly viscous thermal compound with good stability, making it simple to work with while also resisting drips and runs. This paste is very easy to apply and clean up, making it a favorite among system builders.
Best Budget Thermal Paste
3. Noctua NT-H1
Best Budget Thermal Paste
Electrically Conductive: No | Thermal Conductivity: Unlisted | Ease of Use: 4.5 (1=difficult, 5=simple) | Relative Performance: 4.0 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $2.26
Price/performance ratio
Easy to use and clean
None
Noctua’s NH-H1 is a relatively viscous and stable compound that applies easily and allows for easy clean-up. A screw-on cap syringe maintains compound longevity to prevent drying and is priced as an attractive paste for those on a budget. You really can’t go wrong with this paste — it is sufficient for the majority of builds.
4. Gelid GC-Extreme
Runner-Up: Best Budget Thermal Paste
Electrically Conductive: No | Thermal Conductivity: 8.5 W/mk | Ease of Use: 4.0 (1=difficult, 5=simple) | Relative Performance: 3.5 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $3.70
Consistent spread
Value pricing
Easy to use and clean up
Requires some pressure on the syringe to apply correctly
Gelid GC-Extreme is a moderately viscous compound, making it very stable during application and it spreads consistently, although it requires slow consistent pressure on the syringe plunger for correct placement. This TIM is relatively easy to use and simple to clean up.
5. Thermal Grizzly Conductonaut
Best Liquid Metal Thermal Paste
Electrically Conductive: Yes | Thermal Conductivity: 73.0 W/mk | Ease of Use: 2.0 (1=difficult, 5=simple) | Relative Performance: 4.5 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $22.99
For extreme overclocking
High performance
Price
Difficult to apply
Electrically conductive
Thermal Grizzly Conductonaut liquid metal thermal compound has a very thin, runny consistency and is applied to the CPU integrated heat spreader with a syringe delivery system and capillary needle applicator tip. Excess material can be siphoned off using a special extraction tip (by pulling the syringe plunger to withdraw excess liquid metal).
6. CoolLaboratory Liquid Pro
Runner-up: Best Liquid Metal Thermal Paste
Electrically Conductive: Yes | Thermal Conductivity: 79.0 W/mk | Ease of Use: 2.0 (1=difficult, 5=simple) | Relative Performance: 4.5 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $24.94
For extreme overclocking
High performance
Price
Difficult to apply
Electrically conductive
CoolLaboratory Liquid Pro is a liquid metal compound that is applied with a syringe and capillary tip to the CPU integrated heat spreader, and it ships with a scouring pad and alcohol swab for prep and cleanup.
Application and spreading of the liquid metal are made easier with cotton swabs and generous pressure for even distribution. Liquid metal can cause reactions with some metals, such as aluminum, and some reports of reactions with copper are also documented by end-users.
Best Budget Liquid Metal Thermal Paste
7. Phobya Liquid Metal Compound LM
Best Budget Liquid Metal Thermal Paste
Electrically Conductive: Yes | Thermal Conductivity: Unlisted | Ease of Use: 2.0 (1=difficult, 5=simple) | Relative Performance: 4.5 (1=poor, 5=excellent) | Cleanup: lint-free absorbent paper towels, alcohol pads or cotton swabs and alcohol | Price per gram: $9.00
For extreme overclocking
High performance
Relatively inexpensive for a liquid metal paste
Difficult to apply
Electrically conductive
For a compound that costs less than half as much as leading competitors , Phobya’s Liquid Metal Compound LM offers impressive performance, often besting the the CoolLaboratory Liquid Pro or Thermal Grizzly Conductonaut in our tests.
Liquid Metal Compound LM ships as a syringe with a capillary applicator tip. Spreading the compound is made easier with cotton swabs and generous pressure to develop a consistent coating on the CPU IHS. Can cause reactions with aluminum and other metals.
Thermal Paste Performance Rankings
Here’s a breakdown of our recent testing, and we’ll be adding new pastes to this list over the coming weeks and months. A bit further down, you’ll find our original round of testing with 85 different pastes, some of which we’ve revisited in this new performance hierarchy. That should give you a good idea of how other pastes that we haven’t retested yet would compare to the newer pastes.
We’ll go into our breakdown of how we test these thermal pastes below, but the key takeaway here is that a single thermal compound can perform differently based upon several variables, such as what type of cooler and mount you use. To cover all the bases, we’ve tested every thermal paste with three variables:
Low-tension Noctua NH-D15 mounting to model an air cooler without a backplate (like an Intel stock cooler, or push-pin coolers)
High-tension Noctua NH-D15 mounting to model higher-end air coolers with a backplate
High-tension mounting with an all-in-one (AIO) liquid cooler
Below you’ll find the temperatures we measured on our standardized test bench, and as always, lower temperatures are better. We’ve also included value charts in the album below as well.
Image 1 of 6
Image 2 of 6
Image 3 of 6
Image 4 of 6
Image 5 of 6
Image 6 of 6
Thermal Paste Test Results
As shown in the slides above, the outcome of the different testing scenarios resulted in some interesting comparisons among the compounds. We’ll look at the overall thermal load averages for each, as well as the relative performance value of each in terms of price-per-gram.
Air Cooling – Low Tension Mount Liquid metal compounds make their way to the top of the temperature chart with slightly cooler values than the traditional pastes. Interestingly, the difference between the top and bottom of this chart is less than 4C.
When we look at each compound in terms of unit cost-per-gram, liquid metal compounds are generally more costly, making their overall performance value more of a luxury than a necessity. However, Phobya LM bucks this trend as it shows you can get good liquid metal compound performance at bargain pricing.
Air Cooling – High Tension Mount Liquid metals are again at the top of the list, and again, there’s a relatively small 4C delta between the top and bottom of the list. Only a few minor changes to the pecking order have occurred, and these easily fall within the tight margins of standard deviation.
Small differences in temperatures jostle the chart a bit, but overall the more budget-friendly compounds show much more promising performance value due to relatively close load ranges. We see a trend that most good and great thermal compounds perform very similarly, but might be worlds apart in price per gram.
All In One (AIO) Liquid Cooler – High Tension Mount It’s no surprise that the liquid metal compounds once again sit atop the thermal comparison. Still, a 6C difference between all 15 thermal compounds tested shows there isn’t a lot of variation, even with a 360 all-in-one cooler with a push+pull fan setup for performance.
The performance value chart shows most of the liquid metal compounds with a poor value-per-gram rating, but once again, Phobya LM proves to be the outlier for that sub-category with its budget price tag.
We’re also seeing the trend of larger gram-per-package for thermal pastes (2-5g syringes) along with lower price proving to be stiff competition against the more expensive liquid metals (0.5-1.0g syringes) that cost upwards of twice as much, or more.
Key Takeaways Thermal compounds might always be compared and debated over, but the simple fact remains: PC system building needs thermal compounds to effectively dissipate thermal loads. Without them, our beloved gaming and content-producing machines would struggle to keep components cool during heated frag sessions, heavy workstation computations, or just simply browsing the web.
Not every system will need the most expensive compound, so even the most budget-minded system builders can rest easy knowing that even lower-cost pastes can still prove very effective. You can see our historical testing results with far more pastes after the test setup.
For our thermal compound tests, we use the same hardware, overclock, and configuration for each and every test to minimize environment variables in our testing.
We tested each thermal paste with a low-tension air cooler mount, high-tension air cooler mount, and a high-tension AIO liquid cooler mount. Each application was given a 1-hour burn-in using Prime95 with ten load and cooling cycles over the course of the hour; six minutes each with a ten-second cool-down between. Each testing load session was then executed for a one-hour load period, again using Prime95.
For the air cooling low- and high-tension tests, we used a large Noctua NH-D15 air cooler. We created the low tension mount environment by torquing the mounting screws to 1.13 Nm (10 in/lbs).
The low tension mounting tests help simulate cooler installs that might not use a backplate (like push-pin coolers) or those that do not allow high tension and compression across the CPU integrated heat spreader. Also, repeated tests are not consistently possible with push-pin mounted coolers: The pins can degrade after several mount cycles, which meant we needed to simulate these to maintain consistent test results.
Our high-tension air cooler mount involved tightening the mounting screws fully to the mounting plate and shows thermal paste performance with air coolers with backplates that allow for tighter mounting.
We used the EK-AIO Elite 360 D-RGB for the liquid cooling tests with all fans in a push+pull configuration. We performed these tests with the AIO pump block fully tensioned (tightened). We didn’t test the AIO with a low-tension mount because liquid coolers almost always employ a backplate that allows a high-tension mount.
Overall, each compound was evaluated and stressed over a regimen cycle covering six hours with two different coolers and different mounting tensions, making for no less than 90 hours of compound testing for our initial round of tests.
Legacy Thermal Paste Testing – 85 Pastes Tested
Image 1 of 5
Image 2 of 5
Image 3 of 5
Image 4 of 5
Image 5 of 5
Here’s our legacy thermal paste testing chart that we generated back in 2017. As you’ll notice, many of the same popular pastes in 2017 are still on the market, making it into our new round of 2021 testing.
We do have new testing equipment for our updated tests, such as a more modern CPU, motherboard, and coolers for our tests, but we adhered to the same test methodology employed in these prior tests. That means that you can use these legacy tests as a decent approximation of how the newer pastes compare to the older pastes, too.
We’ll be adding more thermal pastes to our new hierarchy; stay tuned.
The cryptocurrency mining business is booming again and, according to a PC Gamer report, Palit may be getting ready to ride the wave. The company has registered multiple P106 graphics cards with the Eurasian Economics Commission (EEC). As always, these registered models may or may not make it to the market.
Palit, like many other brands, launched the P106-100 back in 2017. The graphics card was based on the GP106 Pascal silicon, which is the same die that powers the highly successful GeForce GTX 1060. In effect, the P106-100 and GeForce GTX 1060 were literally the same graphics card, aside from the removal of the video outputs on the first.
As PC Gamer noted, the new part numbers sport a different nomenclature in comparison to the original part numbers. The registered part numbers fall in line with Palit’s last-generation GTX 16-series graphics cards. However, it isn’t likely that the manufacturer will update graphics card to the TU116 Turing silicon. Given the recent revitalization of the cryptocurrency mining business, Palit may just be dumping leftover inventory that it may have in its warehouse.
Nvidia has recycled Turing to launch its latest Cryptocurrency Mining Processor (CMP) lineup of graphics cards. The 30HX is Nvidia’s entry-level crypto-mining SKU. And as we’ve seen, it’s likely based around the TU116 silicon, the die inside the GeForce GTX 1660 Super. Nvidia rates its 30HX with an Ethereum hash rate up to 26 MH/s. According to Minerstat, the P106-100 is good for a hash rate of 21 MH/s, which isn’t bad for a four-year-old graphics card.
The 30HX and 40HX are scheduled to launch in the first quarter of the year. But Nvidia hasn’t revealed the pricing for any of its CMP devices yet. However, we’ve already seen mentions of the 30HX and 40HX in the latest GeForce Game Ready 461.72 WHQL driver, so they should be right around the corner. If Palit prices the new P106-100 just right, the manufacturer could definitely steal some of the cryptocurrency profit pie from Nvidia’s CMP.
Intel’s SSD 670p delivers on both speed and security with faster throughput and hardware encryption support. While Intel’s SSD 670p leverages QLC flash, it is tweaked and tuned where it matters, making it a stellar consumer-oriented PCIe 3.0 NVMe SSD.
For
Solid performance
Large dynamic SLC cache with fast recovery
AES 256-bit hardware encryption
Software package
5-year warranty
Against
High launch MSRP
Slow write speeds after the SLC cache fills
Low endurance compared to TLC SSDs
Features and Specifications
Intel’s SSD 670p rocks an updated PCIe 3.0 x4 controller and the company’s new 144-Layer QLC flash, offering up efficient and performant storage in a slim-yet-spacious M.2 package. It also comes with AES 256-bit hardware encryption support, making it a secure pick for those with sensitive data.
Intel’s SSD 6-series M.2 NVMe SSDs brought Intel’s QLC flash to the mainstream market. With bargain price points and plenty of capacity, both the SSD 660p and SSD 665p have been excellent options for value seekers willing to accept less-than-mainstream levels of performance. However, while responsive, these SSDs couldn’t quite keep pace with the best SSDs on the market and lacked the endurance ratings we have grown accustomed to with TLC flash-based SSDs.
While those SSDs have historically been good budget picks, Intel’s SSD 670p aims to take the SSD 6 series to another level. Intel’s SSD 670p is more than a simple revision — the low-cost M.2 SSD comes with the latest leading-edge NAND technology, improved endurance, and plenty of optimizations for office productivity and gaming.
Specifications
SSD 670p 512GB
SSD 670p 1TB
SSD 670p 2TB
Pricing
$89.00
$154.00
$329.00
Capacity (User / Raw)
512GB / 512GB
1024GB / 1024GB
2048GB / 2048GB
Form Factor
M.2 2280
M.2 2280
M.2 2280
Interface / Protocol
PCIe 3.0 x4 / NVMe 1.3
PCIe 3.0 x4 / NVMe 1.3
PCIe 3.0 x4 / NVMe 1.3
Controller
Silicon Motion SM2265
Silicon Motion SM2265
Silicon Motion SM2265
DRAM
DDR3L
DDR3L
DDR3L
Memory
Intel 144L QLC
Intel 144L QLC
Intel 144L QLC
Sequential Read
3,000 MBps
3,500 MBps
3,500 MBps
Sequential Write
1,600 MBps
2,500 MBps
2,700 MBps
Random Read (QD1)
20,000 IOPS
20,000 IOPS
20,000 IOPS
Random Write (QD1)
54,000 IOPS
54,000 IOPS
54,000 IOPS
Random Read
110,000 IOPS
220,000 IOPS
310,000 IOPS
Random Write
315,000 IOPS
330,000 IOPS
340,000 IOPS
Security
AES 256-bit FDE
AES 256-bit FDE
AES 256-bit FDE
Endurance (TBW)
185 TB
370 TB
740 TB
Part Number
SSDPEKNU512GZX1
SSDPEKNU010TZX1
SSDPEKNU020TZX1
Warranty
5-Years
5-Years
5-Years
The SSD 670p is available in capacities of up to 2TB with pricing that ranges from $0.15-$0.17 per gigabyte. Intel tweaked and tuned the drives for low queue depth requests, making them snappy in everyday desktop PC tasks, and also focused on tuning for mixed read/write workloads to assure strong performance in more demanding workloads.
The SSD 670p can deliver up to 3.5/2.7 GBps of sequential read/write throughput and sustain up to 20,000/54,000 random read/write IOPS at a queue depth (QD) of 1, an important metric to quantify snappiness during non-demanding tasks. Peak performance reaches up to 310,000/340,000 random read/write IOPS at a queue depth of 256. Notably, these specifications are based on the SSD’s dynamic SLC cache. Because cache performance is so important, Intel optimized the design to improve performance when the drive is nearly full.
Like the SSD 665p, the 2TB SSD 670p’s cache measures up to 280GB when the device is empty, but the dynamic cache remains available until the drive is 85% full, an improvement over the 75% threshold with the older drive. At that point and beyond, the drive will operate with only a static SLC cache that measures 6GB per 512GB of capacity.
Intel’s SSD 670p features global wear-leveling and supports robust LDPC error correction capabilities, end-to-end data path protection, and DRAM ECC and SRAM ECC for data reliability. These, in conjunction with the new flash’s enhancements, enable Intel to back the Intel SSD 670p with a five-year warranty and improve its write endurance over its predecessor. The 670p’s write endurance rating is 185 TBW for every 512GB of capacity, but the drive is only overprovisioned by 7% from the factory (2% less than Samsung drives, on average).
While Intel has improved the SSD 670p’s endurance rating over its predecessors (endurance jumps from 100-150TBW per 512GB to 185TBW), the drive still can’t entirely overcome the endurance penalties associated with QLC flash. For instance, the Adata XPG Gammix S50 Lite comes with TLC flash and features 370TBW per 512GB of capacity, while the TLC-powered Samsung 970 Evo Plus comes with 300TBW per 512GB of capacity. The SSD 670p trails these drives in endurance but most consumers will write roughly 60-160TB of data in five years, meaning the 670p should provide plenty of endurance for a typical user.
Software and Accessories
Image 1 of 3
Image 2 of 3
Image 3 of 3
Intel supports the SSD 670p with the Intel Memory and Storage Tool (Intel MAS). This software allows you to monitor drive health, run diagnostic tests, update the firmware, and manually clear the SLC cache. Additionally, Intel provides an NVMe driver for its consumer SSDs, but at the time of publication, the latest downloadable version (version 5.1.0.1003) will not install on the SSD 670p.
A Closer Look
Image 1 of 3
Image 2 of 3
Image 3 of 3
Intel’s SSD 670p comes in an M.2 2280 single-sided form factor for compatibility with the latest thin and light devices. Aesthetically, the SSD 670p’s green PCB and the white sticker aren’t the most attractive, but that only matters if you’re going to place the drive inside a new desktop build with a see-through panel and leave it exposed without a heatsink.
Image 1 of 2
Image 2 of 2
The SSD 670p uses a custom Silicon Motion’s SM2265 controller, one of the company’s latest NVMe SSDs controllers specifically optimized for use with Intel’s latest 144-Layer 3D QLC flash. The SM2265’s architecture is similar to the SM2267 that powers the Adata XPG Gammix S50 Lite, but it lacks the higher-bandwidth PCIe Gen4 interface, employing a Gen3 PHY instead.
The SM2265 leverages two Arm Cortex R5 CPU cores for I/O processing and interfaces with a single 2Gb Nanya DDR3L-1866 DRAM chip to buffer FTL metadata. It’s produced on a 28nm process node for cost-effective production and cool thermals. The drive also supports power-saving features, including ASPM, APST, and L1.2 sleep (rated at just 3mW of consumption), along with hardware-accelerated AES 256-bit encryption with support for Pyrite 2.0 for the security-conscious.
Our 2TB sample comes armed with sixteen dies of Intel’s latest 144-Layer QLC flash, with eight dies per package. These dies interface with the controller at frequencies up to 1,200 MTps over four flash channels, up from 667 MTps with Intel’s previous-gen flash. Part of that improvement stems from scaling to the highest bit densities (13.8 Gb/mm^2) and layer heights the company has ever achieved. To reach 144 layers, Intel has moved to a 3-deck stack (48+48+48 layers) design, an industry first for mass-produced flash.
In this block-by-deck architecture, Intel separates each deck with a dummy layer, and each deck can operate as either SLC or QLC. For more efficient block erases, each deck can be erased without touching the data stored on the other decks. Intel says this approach helps tremendously with garbage collection and boosts quality of service (QoS) significantly.
The flash architecture also incorporates CMOS under the array (CuA) tech to boost density and features quad-plane access to enhance parallelism. Intel uses a floating gate cell design to optimize die space and boost data retention. Intel also claims the design can handle more parallel data operations.
Image 1 of 3
Image 2 of 3
Image 3 of 3
Intel also implemented a new independent multi-plane read operation (IMPRO) technique to double the number of read operations by splitting the four planes into dual two-plane groups that can be read asynchronously, allowing the SSD to simultaneously read from the TLC and QLC portions of the flash. The asynchronous nature of IMPRO can create noise coupling. To mitigate that effect, Intel configures the charge-pumps, wordline/bitline regulators, and drop-out (LDO) regulators to drive separate loads in each plane group.
Additionally, to address the sensitivity of QLC technology and reduce first-pass program penalties, the flash features a 4-16 multi-pass programming algorithm and a 1-2-6-6 Gray code. The cells are optimized via a quad-level dynamic start technique that first programs the cell to a 4-level state, and then after the data is read from the cells, the cells are programmed into the final 16-state level.
The MSI GP66 Leopard is a powerhouse gaming notebook housed in a fairly subtle shell that also boasts a comfortable keyboard. It also offers plenty to upgrade or repair, but it’s a chore getting inside.
For
Strong gaming performance
Comfortable keyboard
Replaceable components
Subtle design for a gaming notebook
Against
Difficult to open
Touchpad feels cheap
Too much bloatware
Just because you grow up a bit doesn’t mean you need to stop having fun. The MSI GP66 Leopard ($1,799.00 to start, $2,599.00 as tested) is a powerhouse gaming notebook with an Nvidia GeForce RTX 3080 and Intel Core i7-10870H. But it would fit in anywhere, whether it be an office or a gaming room, thanks to its matte-black stylings. If you’re looking for gaming power without all of the flash, this might be on your list of the
best gaming laptops
.
If you ran some workstation tasks on this, you might believe it was a work machine. Only when you turn on the RGB keyboard do you know it’s time to play. There aren’t red stripes, or, say, an RGB lightbar like MSI’s other models.
It’s a powerhouse, and many of the components are upgradeable for replacement down the line. But while the GP66 Leopard is all grown up, there are still some areas, like its touchpad and its bloatware, where it needs some more maturing.
Design of the MSI GP66 Leopard
Image 1 of 2
Image 2 of 2
As far as MSI’s gaming notebooks go, the Leopard is quite plain. While some of its other laptops have red accents or RGB light strips, the GP66 is an all-black affair. On the aluminum lid, even the dragon shield logo is tone-on-tone. Only the hinges, which are connected to aggressively shaped plastic molds, suggest this is anything other than a workstation PC.
The only real sign of the GP66’s gaming prowess is the keyboard, which has RGB backlighting courtesy of SteelSeries. But the black aluminum deck and the fairly thin bezels around three sides of the display suggest just a premium notebook. The bottom cover is plastic.
Image 1 of 3
Image 2 of 3
Image 3 of 3
The Leopard is a thick machine at 0.92 inches tall, so there’s plenty of room for ports. However, MSI has placed only a few of them on the sides: The right side has a pair of USB 3.2 Gen 1 Type-A ports, while the left side has another Type-A and the headphone jack. The rest of the ports — USB 3.1 Gen 2 Type-C, HDMI, RJ-45 Ethernet and the charging port — are all on the rear of the device. Alienware has been doing this for years, and it’s a benefit if you use your laptop like a desktop replacement and don’t move it around much. But it can also be inconvenient if you like to use your laptop on your lap.
At 14.09 x 10.51 x 0.92 inches and 5.25 pounds, the Leopard isn’t exactly portable in the toss-it-in-a-bag sense. Dell’s 15-inch competitor, the Alienware 15 m4, is slightly lighter at 5 pounds and a similar size 14.2 x 10.9 x 0.9 inches, but its design is slightly more sleek. The Gigabyte Aorus 17G is expectedly larger with a bigger screen, at 5.95 pounds and 14.9 x 10.8 x 1 inches.
3x USB 3.2 Gen 1 Type-A, USB 3.2 Gen 2 Type-C, HDMI, 3.5 mm headphone jack, RJ-45 Ethernet
Camera
720p
Battery
65 Wh
Power Adapter
230W
Operating System
Windows 10 Home
Dimensions(WxDxH)
14.09 x 10.51 x 0.92 inches / 357.89 x 266.95 x 23.37 mm
Weight
5.25 pounds / 2.38 kilograms
Price (as configured)
$2,599.00
Gaming and Graphics on the MSI GP66 Leopard
MSI opted for a powerful implementation of the Nvidia GeForce RTX 3080 for the GP66 Leopard. This version has a 130W max graphics power and a
boost clock
of 1,605 MHz.
Besides running benchmarks, I tried playing Control, a game I use often on laptops with RTX GPUs because of how tough it is to run and because
ray tracing
has a truly noticeable effect. With the settings maxed out at 1080p and with ray tracing on high, the game ran between 52 and 57 frames per second as I traded shots with hiss guards surrounding a control point, though it went as high as 70 during exploration.
Image 1 of 5
Image 2 of 5
Image 3 of 5
Image 4 of 5
Image 5 of 5
On Shadow of the Tomb Raider (1080p, highest), the GP66 hit 106 frames per second. The Aorus 17G, with an RTX 3080
Max-Q
ran the game at 86 fps, while the Alienware m15 R4 with an RTX 3070 played it at 77 fps.
The Leopard played Grand Theft Auto V (1080p, very high) at 125 fps, beating the Aorus (100 fps) and Alienware m15 (108 fps).
MSI’s laptop ran Far Cry New Dawn (1080p, ultra) at 103 fps, beating both the Alienware and Aorus by over 10 frames per second.
The GP66 Leopard outperformed on Red Dead Redemption 2 (1080p, medium), playing at 82 fps.
It also won out on Borderlands 3 (badass, 1080p), at 99 fps, while the Aorus 17G ran at 79 fps and the Alienware hit 84 fps.
We also ran our gaming stress test on the GP66 Leopard by looping the Metro Exodus benchmark at RTX settings for 15 runs, simulating roughly half an hour of gaming. It ran at a largely steady average of 76.38 frames per second across the runs. The CPU ran at an average of 3.73 GHz and an average temperature of 61.85 degrees Celsius (143.3 degrees Fahrenheit). The GPU ran at an average of 1.1 GHz and 61.49 degrees Celsius (142.68 degrees Fahrenheit).
Productivity Performance on the MSI GP66 Leopard
Beyond gaming, the GP66’s Intel Core i7-1070H and GeForce RTX 3080, along with 32GB of RAM and a 1TB SSD are powerful parts that should aid in creative endeavors like streaming or video editing.
Image 1 of 3
Image 2 of 3
Image 3 of 3
On the Geekbench 5 overall performance benchmark, the GP66 earned a single-core score of 1,274 and multi-core score of 7,848. The Gigabyte Aorus 17G had scores of 1,265 and 7,895, respectively, while the Alienware 15 R5 notched scores of 1,252 and 7,642.
On our file transfer test, the Leopard copied and transferred 25GB of files at a rate of 1,059.78 MBps, falling just below the Alienware but ahead of the Aorus.
It took the GP66 Leopard seven minutes and three seconds (7:03) to complete our Handbrake test, transcoding a 4K video to 1080p. That’s slightly faster than the Alienawre (7:07) and far speeder than the Aorus (8:33).
Display on the MSI GP66 Leopard
Our review unit came equipped with a 15.6-inch, 1920 x 1080 (FHD) display with a 240 Hz refresh rate. The most demanding titles won’t run that fast, but you can take advantage of it if you like playing esports titles like Overwatch, Fortnite or Rocket League.
When I watched the trailer for the upcoming Mortal Kombat movie, I felt the need to turn up the brightness for the best experience. Cole’s yellow gloves popped, as did Kano’s red eye lasers against a dark background, but the screen was largely serviceable rather than special.
When I played Control, the screen was bright enough, even in some dark spaces. That game has a lot of red, and it really popped, especially against the Oldest House’s dark walls.
MSI’s panel covers 78.5% of the DCI-P3 color gamut, matching what we saw on the Aorus 17G. We reviewed the Alienware m15 R4 with a 4K
OLED
screen, so it’s not surprising to see superior coverage there.
However, at 277 nits of brightness, the screen was dimmer than both the Aorus (300 nits) and the Alienware (362 nits).
Keyboard and Touchpad on the MSI GP66 Leopard
MSI’s inputs are a mixed bag. Arguably the more important part of a gaming laptop, the keyboard, is the superior of the two. For years now, MSI has partnered with SteelSeries on its keyboards, and they’ve typically been quite good. On the Leopard, the keys are a bit more shallow than I would like, but they’re still fairly comfortable to type on. I hit 120 words per minute with a 2 percent error rate, which is about as fast as I ever get on the 10fastfingers.com typing test.
The 2.5 x 4.1-inch touchpad is fine for navigating and gestures with its Windows precision drivers, but it felt like cheap plastic compared to the aluminum around it. On top of that, I found I had to click harder than on most laptops. For gaming, you should be using a mouse anyway, but this could be a little better for general productivity use.
Audio on the MSI GP66 Leopard
Image 1 of 2
Image 2 of 2
When I listened to Daft Punk’s “Something About Us,” the various synths and samples were quite clear through the laptop’s bottom-firing speakers. The vocals, purposefully distorted, still stood out on top of the track, along with some piano backing. The low end, including some drums, could have used a bit more punch.
There is a semblance of bass, which many laptops don’t offer, and I was able to tune it more to my liking in the Nahimic audio software.
When I played Control, Jesse’s internal thoughts were clear, especially as her narration ran over the creeping voices of the hiss. At some points, though, I wished I was able to make the volume louder, though this is something else that could be solved with headphones.
Upgrading the MSI GP66 Leopard
The GP66 Leopard is pretty easy to upgrade or repair. Well…it is once you manage to get inside. Compared to some other gaming notebooks, it’s like breaking into Fort Knox.
At first, eleven screws separate you from the GP66 Leopard’s internals. A size 000 Phillips head screwdriver will do the trick. Note, though, that one screw is beneath a factory seal, which is a questionable practice at best. If you send this back through warranty, MSI will know you opened it.
Even after removing all the screws, the system was too tight to open at first. No spudger or pick would fit in the cracks in the chassis. I tried this for longer before I cared to admit, before I eventually found
a YouTube video
from someone that had cracked the case. There is a decorative cover around the hinges, which you can pop off with a spudger. Once that’s off, you can slowly move your way around from the rear ports to the front of the case and carefully remove the bottom.
Once you’re in there, you’ll find that the RAM, Wi-Fi card and the SSD are replaceable. There are two PCIe
m.2 SSD
slots, and since ours came with a sole 1TB boot drive, there is room to expand. The 65 Wh battery, too, is replaceable.
Battery Life on the MSI GP66 Leopard
The MSI’s GP66 Leopard’s 65W hour battery isn’t going to last it terribly long on a charge. This, unfortunately, is a trend on gaming notebooks, but the Leopard with its full-power RTX 3080, didn’t last as long as competitors.
MSI’s notebook endured for two hours and 25 minutes on our test, which browses the web, runs OpenGL tests and streams video over Wi-Fi, all at 150 nits of brightness. The Alienware m15 ran for 4:01 and the Gigabyte Aorus 17G ran for 4:42.
Heat on the MSI GP66 Leopard
Call it the MSI GP66 Jet Engine.
To keep its components cool, the GPU fans run hard and loud (especially in extreme performance mode, which MSI sent the GP66 Leopard to us set to by default). Admittedly, if you use headphones this is a bit less of a problem.
We took surface temperatures while running our Metro Exodus gauntlet (see the gaming performance section above).
Image 1 of 2
Image 2 of 2
The center of the keyboard, between the G and H keys, measured 38.6 degrees Celsius (101.48 degrees Fahrenheit), while the touchpad reached 25.6 degrees Celsius (78.08 degrees Fahrenheit). The hottest point on the bottom of the laptop was 46.7 degrees Celsius (116.06 degrees Fahrenheit).
Webcam on the MSI GP66 Leopard
MSI has a laptop with a 1080p
webcam
coming this year. This isn’t it.
No, the GP66 Leopard still has a 720p webcam, and an image at my desk was grainy, though at least it was color accurate with my blue eyes and green t-shirt.
One minor annoyance is that the light that notifies you the camera is on blinks, rather than staying on and static. This is extremely distracting when you’re having a video call or streaming and want to focus on what’s on the screen.
Software and Warranty on the MSI GP66 Leopard
There is quite a lot of software preinstalled on the GP66, which has been a trademark of MSI laptops for a bit now. Unfortunately, a lot of it is bloatware.
Let’s start with the good stuff (it’s a shorter list). There’s MSI Dragon Center, which lets you monitor CPU and GPU usage and other stats, as well as change between different modes of performance. SteelSeries Engine 3 lets you configure the lighting on the keyboard, though I feel MSI should roll this into the other app. Nahimic lets you customize audio profiles.
Aside from that, MSI has added a ton of extra bloat, including the Cyberlink suite (AudioDirector, ColorDirector, PhotoDirector and PowerDirector), as well as Microsoft Sudoku, LinkedIn, Music Maker Jam and Norton Security.
That’s on top of the regular
Windows 10
inclusions, like Roblox, Hulu, Hidden City: Hidden Object Adventure and Adobe Photoshop Express.
MSI sells the GP66 Leopard with a one-year warranty.
Configurations
We tested the MSI GP66 Leopard with an Intel Core i7-1070H, Nvidia GeForce RTX 3080, 32GB of RAM, a 1TB M.2 NVMe PCIe SSD and a 15.6-inch, FHD display with a 240 Hz refresh rate. All of that adds up to a grand total of $2,599. (We have, however, seen an identical model with 32GB of RAM going for $2,499, so be sure to shop around).
For $1,799, you can get the Leopard with a Core i7-10750H, 16GB of RAM, a 512GB SSD and an FHD 144 Hz display.
Bottom Line
If you’re looking for a powerful gaming notebook that draws attention to games, not itself, the GP66 Leopard is worth looking at. The combination of a full-power RTX 3080 and 10th Gen Intel makes for a potent, if loud, gaming machine. And MSI has put it in a chassis that looks and feels fairly adult, even compared to something like the Alienware m15, which has also gone minimalistic.
There are some things MSI needs to work on here: primarily, the touchpad, which feels like an afterthought, and the sheer amount of bloatware that the company includes on its laptops. If you prefer a premium experience, the Alienware may be a better way to go, but a similarly specced model (with a 300 Hz compared to 240 Hz on the Leopard) is a little more expensive as of this writing.
But if you want a gaming notebook with powerful graphics performance, subtle styling and replaceable parts (even if it takes a bit of work to get to them), this Leopard will impress.
One of the world’s most popular and innovative computers officially launched on February 29, 2012. Nine years and 38 million units sold later, the Raspberry Pi powers a huge community of makers, students and businesses. What started as a small project, meant to increase applications for Cambridge University’s computer science program has become a global movement.
Every serious tech enthusiast should own at least one Raspberry Pi or, if you’re like me, 30 Raspberry Pis. You can use one as a lightweight PC, a retro arcade machine or to power a variety of projects, from home security cameras to tic-tac-toe-playing robots, self-driving trash cans or streaming media servers.
In honor of the Raspberry Pi’s ninth birthday, here are 9 key facts about it.
1. Raspberry Pi’s original target was just 1,000 units.
The Raspberry Pi was originally developed to solve a very limited problem: the decreasing number of people applying to study computer science at Cambridge University. The number of applications had dropped from 600 to 250 per year and Eben Upton, who was the director of studies and responsible for admission, became concerned that not enough kids were taking an interest in computers. By providing a low-cost, hackable computer to just a few kids in the UK, Upton intended to get more and better students for his program.
“The stuff we were designing, the business model side, they were all scaled around the idea that if you could get 1,000 units built and into the hands of the right 1,000 kids [you’d solve the problem],” he told us in 2019.
After interest in the project swelled, Upton and his team decided that they needed to make a lot more than 1,000 units. And the target audience expanded from U.K. children to people of all ages in all countries.
2. There have been at least 21 Different Models of Raspberry Pi.
Since its launch in 2012, the Raspberry Pi Foundation has released 20 additional models, 52 if you count all four RAM capacities of the Raspberry Pi 4 B and all 32 variants of the Raspberry Pi Compute Module 4 (which comes in different RAM and storage capacities and with or without Wi-Fi).
The original Raspberry Pi, the Model B had just 256MB of RAM and a single-core, 700 MHz processor. The current mainstream model, the Pi 4 B, has up to 8GB of RAM, a quad-core 1.5-GHz CPU and USB 3.0 ports.
Model
Year Released
CPU
RAM
Raspberry Pi 1 B
2012
700 MHz Broadcom BCM2835 (1 core)
256MB
Raspberry Pi 1 A
2013
700 MHz Broadcom BCM2835 (1 core)
256MB
Raspberry Pi 1 A+
2014
700 MHz Broadcom BCM2835 (1 core)
512MB
Raspberry Pi 1 B+
2014
700 MHz Broadcom BCM2835 (1 core)
512MB
Compute Module 1
2014
700 MHz Broadcom BCM2835 (1 core)
512MB
Raspberry Pi 2 B
2015
900 MHz Broadcom BCM2836 (4 cores)
1GB
Raspberry Pi Zero 1.2
2015
1 GHz Broadcom BCM2835 (1 core)
512MB
Raspberry Pi 2 B v1.2
2016
1 GHz Broadcom BCM2835 (1 core)
512MB
Raspberry Pi 3 B
2016
1.2 GHz Broadcom BCM2837 (4 cores)
1GB
Raspberry Pi Zero 1.3
2016
1 GHz Broadcom BCM2835 (1 core)
512MB
Raspberry Pi Zero W
2017
1 GHz Broadcom BCM2835 (1 core)
512MB
Compute Module 3
2017
1.2-GHz Broadcom BCM2837 (4 cores)
1GB
Compute Module 3 Lite
2017
1.2-GHz Broadcom BCM2837 (4 cores)
1GB
Raspberry Pi 3 B+
2018
1.2-GHz Broadcom BCM2837 (4 cores)
1GB
Compute Module 3+
2019
1.2-GHz Broadcom BCM2837B0 (4 cores)
1GB
Compute Module 3+ Lite
2019
1.2-GHz Broadcom BCM2837B0 (4 cores)
1GB
Raspberry Pi 4 B
2019
1.5-GHz Broadcom BCM2711
1 / 2 / 4 / 8GB (2020)
Compute Module 4
2020
1.5-GHz Broadcom BCM2711
1 / 2 / 4 / 8GB
Raspberry Pi 400
2020
1.8-GHz Broadcom BCM2711
4GB
Raspberry Pi Pico
2021
133-MHz RP2040
264K
Launched in late 2020, the Raspberry Pi 400 is Raspberry Pi’s first standalone computer. Instead of a bare board, the Pi 400 is a keyboard with the equivalent of a Pi 4 inside, though the CPU runs at 1.8 rather than 1.5 GHz. It’s sold either by itself or in a kit with a mouse, power supply, cables and an official guide.
The Raspberry Pi Pico is a completely different type of Pi, a microcontroller board that’s more like an Arduino than a traditional Pi (more on that below).
3. Raspberry Pi Pico marks a new chapter in Pi history.
Released in January 2021, the Raspberry Pi Pico is the company’s first microcontroller and marks the debut of its first custom silicon, the RP2040 CPU. Where every prior Pi has been a full-fledged computer that boots into a, typically Linux, operating system, the $4 Pico falls into the same category as Arduino. The Pico is great for controlling motors, lights and sensors and runs a program as soon as you turn it on, without the need to boot up or the worry that you’ll lose data if you pull the plug without doing a safe shutdown.
The Pico has three built-in ADC (analog-to-digital converters), something that other Pis lack, which allow you to connect directly to analog devices such as joysticks, potentiometers and distance sensors. See our articles about Raspberry Pico vs Arduino and which Raspberry Pi is right for you for more detailed comparisons between the Pico, the Arduino and other Raspberry Pis.
The RP2040 chip which powers the Pico marks an even bigger evolution than the board itself. This 133-MHz, dual-core Cortex M0+ CPU is designed by the Raspberry Pi Foundation and licensed out to other vendors such as Pimoroni and Adafruit who are building an entire ecosystem of RP2040-powered microcontrollers themselves. Even Arduino is getting in on the act, releasing the Arduino Nano RP2040 Connect, which will have built-in Wi-Fi and Bluetooth.
4. The most popular Pi is . . .
The Raspberry Pi 3 B is the best-selling Pi model of all-time, with 13.2 million units sold. Of the various Raspberry Pi 4 B RAM capacities, the 4GB model accounts for approximately twice as many sales as the 8GB and 2GB SKUs.
The Pico has only been on the market for a few weeks, but already has sold 250,000 units with 750,000 on back order. 2020 was the top-selling year for Pi, with 7.1 million units sold.
5. The Pi has more than 20 operating systems.
Raspberry Pi OS, a customized version of Debian, is the official operating system for Raspberry Pi boards, but it’s not the only choice, by far. From Pi-friendly versions of Ubuntu and FreeBSD Linux to unofficial ports of Android and Chrome OS, there are a ton of different operating systems available. Some of the most interesting include:
LibreELEC: A lightweight OS that runs the Kodi open-source media software.
RISC OS: Designed for ARM chips, this unique OS promises faster speeds than Linux.
Chromium OS: Turn your Pi into a Chrome Box.
Windows 10: Yes, you can install Windows 10 on a Raspberry Pi 4 (with some hacks), but it’s so slow you probably won’t want to keep using it.
Lakka: Specially designed for retro gaming, this OS comes with a series of emulators pre-loaded.
6. There are two Raspberry Pis in space.
The International Space Station is home to two “Astro Pis,” which are specially modified Raspberry Pi B+ models that have been “space hardened” and equipped with the official Raspberry Pi Sense HATs . The European Space agency runs periodic contests where children submit code to be run on the devices.
Though the Pis are older models, they recently got a storage upgrade as the Astro Pi project sent 256GB microSD cards to the ISS to replace the 32GB cards they were using.
7. The Raspberry Pi Foundation wants products to live forever.
Even though the Raspberry Pi B+, Raspberry Pi 2, Raspberry Pi 3 B and other models are way out of date, the Raspberry Pi Foundation continues to actively manufacture and sell them. The organization hates to End-of-Life (EOL) products, because there are industrial clients who may still need them even if they are really old.
“EOLing product is death. We’ve EOLed [just] five products in our life,” Upton told us in 2019. He said that the only five products that Raspberry Pi discontinued include the Pi 1A and Pi 1B, because “the Pi 1A+ and B+ are a better implementation of that world.”
Even after the price of the Raspberry Pi 4 B (2GB) dropped to $35, the 1GB model remained available at the same price, because some customers may still want it as a drop-in replacement.
8. The price of Raspberry Pi has dropped, relative to inflation.
The original Raspberry Pi cost $35 in 2012 while the Raspberry Pi 4 B (2GB) costs the same price today. However, if you consider inflation, $35 from 2012 is actually equivalent to $39.88 today. For that same price, you get:
A 40x faster processor (700 MHz, single-core vs 1.5-GHz quad core)
8x the RAM (256GB vs 2GB)
Wi-Fi vs no-Wifi
Dual monitor output vs single HDMI out
USB 3.0 ports vs USB 2 only
If $35 still seems like a lot of money, there are cheaper Pi models. The Raspberry Pi Zero goes for just $5 while the Raspberry Pi Pico, a microcontroller board, goes for a mere $4.
9. There’s a Pi War every year.
Yes, you can compete with your Raspberry Pi. Pi Wars is an annual robotics competition where all of the gear must be built with your choice of Raspberry Pi. Teams at Pi Wars compete in challenges such as navigating obstacle courses, popping balloons and navigating a maze.
There are both human-driven and automated challenges. The next Pi Wars will take place in July 2021.
sRGB stands for Standard Red Green Blue and is a color space, or a set of specific colors, created by HP and Microsoft in 1996 with the goal of standardizing the colors portrayed by electronics. sRGB is the most popular color space used today and the one used for Windows, most web browsers and most console and PC games, unless they’re HDR.
When looking for the best gaming monitor, TV or other electronic, you may notice a product claiming to cover a certain percentage of the sRGB color space. This tells you how much of the sRGB color space the monitor is supposed to be able to reproduce (how accurately it does so is another story that can only be confirmed via testing).
A monitor or other device’s color gamut tells you which color space or spaces the device can reproduce and how much of said space(s) it can portray from 0-100% or even greater. Besides sRGB, other common color spaces include Adobe RGB and DCI-P3, which are both larger — or encompass more colors — than sRGB.
sRGB Standard
The International Electrotechnical Commission (IEC) standardized sRGB in 1999. You can find the official sRGB standard via the IEC here, and it’s also viewable here.
A color space is definable by its triangle on the CIE 1931 XY Chromaticity Diagram created by the International Commision on Illumination (CIE).
The CIE 1931 XY Chromaticity Diagram below shared via BenQ represents the full range of colors the human eye can see. Within that range of colors is a white triangle, which outlines the colors the sRGB standard entails. If you tried to look at an image with colors outside of that triangle on an sRGB monitor, for example, those other additional colors would look inaccurate and undersaturated. Other popular color spaces are also defined in the diagram below.
Color spaces define their range of colors via specific coordinates on the chromaticity chart above. Coordinates for sRGB colors are based on the (International Telecommunication Union Radiocommunication Sector’s (ITU-R’s) BT.709 standard [PDF], which is also known as Rec .709 and ITU 709 and was created for HDTVs.
Here are the color coordinates for the sRGB space:
X Coordinate
Y Coordinate
Red
0.64
0.33
Green
0.3
0.6
Blue
0.15
0.06
White point, also known as D65 or CIE Illuminant D65
0.3127
0.329
sRGB uses 8-bit color (as opposed to 10-bit color or greater). The sRGB standard also uses a gamma value of 2.2.
sRGB vs DCI-P3
While sRGB is the standard, other color spaces can be desirable. For example, Adobe RGB, while not an international standard, is popular in professional environments as it offers a broader range of colors and was made by Adobe, bringing obvious appeal to users of software like Adobe Photoshop. DCI-P3 is another popular space that’s similar to Adobe RGB in terms of the colors it covers but targets video and is internationally standardized.
DCI-P3 and Adobe RGB are similar and both larger (covering more colors) than sRGB. If accurate, a DCI-P3 monitor has a 25% wider color gamut than an sRGB monitor.
Another difference between sRGB and DCI-P3 is that the latter can handle 10-bit color, key for HDR content.
Speaking of HDR, DCI-P3 is also the color space HDR content uses, so if you want the best HDR monitor or display, you should look for something with significant DCI-P3 coverage. Full sRGB coverage wouldn’t do the format complete justice. If you watch HDR content on an sRGB monitor, (assuming it supported the HDR format), you wouldn’t be accessing the full range of colors available in that media. But if you want a good HDR screen that also allows you to view Windows and SDR content accurately, you should ensure that the display also has an accurate sRGB mode, where it properly shifts the color gamut to sRGB’s range of colors.
We’ve found that for SDR gaming and entertainment, offering more color than sRGB calls for — whether by oversaturating the sRGB color space or using DCI-P3 — can be enjoyable, even if it’s not entirely accurate. In fact, many screens we test these days offer more than 100% coverage of sRGB, making content made in the sRGB color space look more saturated than perhaps the content creators intended.
This article is part of the Tom’s Hardware Glossary.
Further reading:
Display Testing Explained: How We Test PC Monitors
Best Gaming Monitors
Best Budget 4K Monitors
PC Monitor Buying Guide
We use cookies on our website to give you the most relevant experience. By clicking “Accept”, you consent to the use of ALL the cookies.
This website uses cookies to improve your experience while you navigate through the website. Out of these, the cookies that are categorized as necessary are stored on your browser as they are essential for the working of basic functionalities of the website. We also use third-party cookies that help us analyze and understand how you use this website. These cookies will be stored in your browser only with your consent. You also have the option to opt-out of these cookies. But opting out of some of these cookies may affect your browsing experience.
Necessary cookies are absolutely essential for the website to function properly. This category only includes cookies that ensures basic functionalities and security features of the website. These cookies do not store any personal information.
Any cookies that may not be particularly necessary for the website to function and is used specifically to collect user personal data via analytics, ads, other embedded contents are termed as non-necessary cookies. It is mandatory to procure user consent prior to running these cookies on your website.