Showing posts with label performance. Show all posts
Showing posts with label performance. Show all posts

11.14.2022

First gaming review of GeForce RTX 4080 - up to 19% faster than GeForce RTX 3090 Ti

First gaming review of GeForce RTX 4080 - up to 19% faster than GeForce RTX 3090 Ti

First gaming review of GeForce RTX 4080 - up to 19% faster than GeForce RTX 3090 Ti

Two days left before the sales of GeForce RTX 4080 graphics gas pedal on Ada Lovelace architecture, and Chinese enthusiasts already started to publish first reviews of novelty in circumvention of NVIDIA ban.
For example, performance data was shared by Big Hardware Player user on Bilibili platform.
Previously, he was one of the first to publish reviews GeForce RTX 3080 Ti and RTX 3070 Ti generation Ampere.Image source: NVIDIAChinese enthusiast did not specify the manufacturer tested GeForce RTX 4080, but pointed out that the graphics card is equipped with a dual BIOS.
In one case, the maximum TDP of the novelty is set at a nominal value of 320 watts.
The second firmware increases the maximum TDP of the graphics card to 450W.
The second BIOS the Chinese enthusiast used to test the power efficiency.
Source images here and below: Bilibili / Big Hardware PlayerThe GeForce RTX 4080's gaming tests used Forza Horizon 5, Cyberpunk 2077, Borderland 3, Total War Troy and Shadow of the Tomb Raider.
Below is the average performance of the new gas pedal against the GeForce RTX 4090 and GeForce RTX 3090 Ti.
In gaming benchmarks, the GeForce RTX 4080 averaged 19% faster than the GeForce RTX 3090 Ti and about 23% slower than the GeForce RTX 4090.
Note that GeForce RTX 4090 and GeForce RTX 3090 Ti were tested on a DDR5-6800 PC, whereas GeForce RTX 4080 were tested on a DDR5-6200 system.
However, in both cases the test platform was based on the flagship Intel Core i9-13900K processor.
GeForce RTX 4080 in Cyberpunk 2077 GeForce RTX 4080 in Forza Horizon 5 GeForce RTX 4080 in Shadow of the Tomb Raider In 3DMark the new GeForce RTX 4080 is 25% faster than GeForce RTX 3090 Ti and 21% slower than the new flagship GeForce RTX 40 series.
Performance of GeForce RTX 4080 in synthetic 3DMark testsThe most interesting detail of an early review of GeForce RTX 4080 graphics gas pedal were the results of energy efficiency measurements.
For this test Chinese reviewer used BIOS with inflated to 450W TDP value.
Power consumption of GeForce RTX 4080 in synthetic benchmarksTests showed that the card is close to its maximum performance already at 280W.
GeForce RTX 4080 gaming power consumptionGaming power consumption of GeForce RTX 4080 can be reduced even further, to nearly 260W.
Increased power consumption does not bring gas pedal significant performance boost, at least in tests in 4K resolution.Video review of GeForce RTX 4080 by Chinese enthusiast can be found here.
The video card will go on sale on November 16.

11.10.2022

GeForce RTX 4080 compared to RTX 4090 in games - flagship is up to 38.8 % faster

GeForce RTX 4080 compared to RTX 4090 in games - flagship is up to 38.8 % faster

GeForce RTX 4080 compared to RTX 4090 in games - flagship is up to 38.8 % faster

Overclocking the GPU of GeForce RTX 4080 graphics card with an increase in power consumption limit by only 10 W over the reference value increases the performance by up to 9 %.
This is evidenced by a recent leak, which refers to the portal VideoCardz.
In addition to that appeared in the web data on the gaming performance of new items.Image source: VideoCardzThe video card GeForce RTX 4080 tested in synthetic tests 3DMark in three modes: with standard CPU frequency and power consumption, with increased power consumption limit, as well as with increased and frequency, and the limit of consumption.
Results of GeForce RTX 4080 without overclocking (left), with overclocking (center) and with manual overclocking (right) in 3DMark Time Spy Extreme In 3DMark Time Spy Extreme scenario video card performance increased by up to 9%, and in 3DMark Port Royal - to 7%.Results of GeForce RTX 4080 in 3DMark Speed Way and 3DMark Port Royal after manual overclocking The first independent gaming tests GeForce RTX 4080 also appeared on the web.
Recall that the results of laboratory test gaming performance of graphics cards earlier shared by NVIDIA itself.
GeForce RTX 4080 in Cyberpunk 2077 with ray tracing enabled at 4K and 1440p resolution.
Image source here and below: MEGAsizeGPUThe latest GeForce RTX 4080 performance data was shared by one of the users of Chinese forum Chiphell.
The test was conducted in Forza Horizon 5, Call of Duty: Modern Warfare 2, Cyberpunk 2077, Borderlands 3 and Shadow of the Tomb Raider games.
A Ryzen 7 7600X CPU-based PC was used for testing.
GeForce RTX 4080 in Forza Horizon 5 GeForce RTX 4080 in Call of Duty: Modern Warfare 2 GeForce RTX 4080 in Borderlands 3 GeForce RTX 4080 in Shadow of the Tomb RaiderAnother user of the same forum provided performance data on the flagship graphics card GeForce RTX 4090 on an Intel Core i9-12900F-based PC.
With caveats to platform differences, this allows us to compare the two gas pedals in the same games.
The GeForce RTX 4090 outperforms the GeForce RTX 4080 graphics card by between 7.9 and 38.8 percent: Forza Horizon 5 - 147 FPS vs 127 FPS (+15.7 percent);Shadow of the Tomb Raider - 179 FPS vs 129 FPS (+38.8 percent);Red Dead Redemption 2 - 129 FPS vs 99 FPS (+30.3 percent);Cyberpunk 2077 - 111 FPS vs 83 FPS (+33.7 percent);Call of Duty: Modern Warfare 2 - 137 FPS vs 127 FPS (+7.9 percent).
The GeForce RTX 4080 will be available for sale on November 16.
The card is equipped with AD103 graphics processor with 9728 CUDA cores and 16 GB of GDDR6X memory.

10.31.2022

New GeForce RTX 3060 Ti with GDDR6X memory was faster than GeForce RTX 3070 in some games

New GeForce RTX 3060 Ti with GDDR6X memory was faster than GeForce RTX 3070 in some games

New GeForce RTX 3060 Ti with GDDR6X memory was faster than GeForce RTX 3070 in some games

Galax has provided the first gaming performance data on the new version of the GeForce RTX 3060 Ti with GDDR6X memory, which was introduced a few days ago.
The faster GDDR6X provides up to 35% higher memory bandwidth than GDDR6.
This should have a positive impact on gaming performance, especially at high graphics settings.Image source: GalaxThe Galax company will release a GeForce RTX 3060 Ti Plus graphics card with 8 GB of GDDR6X memory.
The manufacturer compared the new product to the older model GeForce RTX 3070, also equipped with 8 GB of memory, but slower GDDR6.
Recall that the GeForce RTX 3070 uses a GPU with 5888 CUDA cores, while the GeForce RTX 3060 Ti is equipped with a GPU with 4864 CUDA cores.
GDDR6 memory of the former provides a bandwidth of 448 GB/s, the second equipped with GDDR6X memory, this figure is 608 GB/s.
According to the video published by Galax, GeForce RTX 3060 Ti with faster memory provides a slightly higher level of performance compared to the GeForce RTX 3070 in a number of games.
The tests were conducted in three games: Spider-Man Remastered, Cyberpunk 2077 and A Plague Tale: Requiem.
Unfortunately, the manufacturer did not specify what resolution and graphics settings were used.
In the first two games, Galax compared two almost identical scenes.
In Spider-Man Remastered, the GeForce RTX 3060 Ti with GDDR6X memory provided 81 FPS of performance, while the GeForce RTX 3070 provided 75 FPS.
In Cyberpunk 2077, the advantage of the model with GDDR6X memory remained.
It provided an average of 70 frames per second, and the GeForce RTX 3070 model - 61 frames per second.
But in A Plague Tale: Requiem game GeForce RTX 3070 model was faster - 81 vs.
78 FPS of GeForce RTX 3060 Ti with GDDR6X memory.Officially the company has not yet presented GeForce RTX 3060 Ti Plus card.
At the time of writing this text the gas pedal is not listed on its official website.
However, Galax's statement once again confirms that this video card model with GDDR6X memory does exist and will soon go on sale.

Intel says Core i9-13900K is on average 11% faster than AMD Ryzen 9 7950X in games

Intel says Core i9-13900K is on average 11% faster than AMD Ryzen 9 7950X in games

Intel says Core i9-13900K is on average 11% faster than AMD Ryzen 9 7950X in games

According to Intel itself, the new Raptor Lake series Core i9-13900K processor is on average 11% faster than the flagship AMD Ryzen 9 7950X in games.
One of the VideoCardz reader shared photos from Intel presentation, anticipating the upcoming release of chips, where in particular the performance of new 24-core flagship processor was concerned.Image source: IntelThe slides showed Intel Core i9-13900K performance was compared directly with the direct competitor Ryzen 9 7950X.
So, the new Intel chip was up to 22% faster in games, compared to AMD's solution.
For example, such a difference can be expected in The Riftbreaker game.
In this case, if we talk about the average difference in performance, it is about 11% in favor of the Intel chip.
Intel Core i9-13900K vs AMD Ryzen 9 7950X in games.
Image source: VideoCardz / IntelOn another slide shown, Intel showed the performance of the Core i9-13900K in various applications.
PugetBench, Procyon, AutoCAD/AutoDesk were selected for the tests.
According to Intel, in half of the tests between the two flagships there is parity in performance.
In Photoshop, Intel's new product is slower by 3% than AMD's processor, while in AutoDesk/AutoCAD design programs, on the contrary, it is 4% and 16% faster.
Intel Core i9-13900K vs AMD Ryzen 9 7950X in software.
Source image: VideoCardz / IntelIt should be noted that the information on what conditions and with what PC components the above tests were conducted, is not yet available.
Intel Core 13th generation processors will be available for sale on October 20.
On the same day, the first independent reviews of these chips should appear.

10.24.2022

NVIDIA: GeForce RTX 4090 graphics card will provide over 500 FPS at 1440p resolution in Overwatch 2

NVIDIA: GeForce RTX 4090 graphics card will provide over 500 FPS at 1440p resolution in Overwatch 2

NVIDIA: GeForce RTX 4090 graphics card will provide over 500 FPS at 1440p resolution in Overwatch 2

There is still a month before the GeForce RTX 4080 is released, but the flagship GeForce RTX 4090 will go on sale next week.
On this occasion, NVIDIA company decided to share the data about the performance of their most powerful card in the game Overwatch 2.
There is information about the performance of other models as well.Image source: VideoCardz According to the published NVIDIA graphics, GeForce RTX 4090 is capable of delivering more than 500 frames per second at 2560 × 1440 resolution at maximum graphics settings in Overwatch 2.The difference between the GeForce RTX 4090 and GeForce RTX 4080 16GB is very significant - 139 frames per second.
Both cards are positioned by the manufacturer as solutions that can deliver more than 360 frames per second, which is ideal for owners of monitors that support 360 Hz refresh.
Image source: NVIDIAWhat about the younger version of GeForce RTX 4080, equipped with 12 GB of memory, it in the same game Overwatch 2 demonstrates just under 300 frames per second.
Unfortunately, the GeForce RTX 3090 has not been added to the performance chart for comparison.
However, NVIDIA did provide information on the performance of the GeForce RTX 3080, which the GeForce RTX 4080 12GB beats by 47 frames per second.NVIDIA recommends the following graphics settings for Overwatch 2 depending on which graphics card you have:NVIDIA GeForce RTX 3060 - 1080p to get 144+ FPS;NVIDIA GeForce RTX 3080 Ti - 1080p to get 360+ FPS;NVIDIA GeForce RTX 4080 16GB - 1440p to get 360+ FPS;NVIDIA GeForce RTX 4090 - 1440p to get 500+ FPS.

10.01.2022

You can play TES: Skyrim, Battlefield V and many other games on the Ryzen 7000 integrated graphics

You can play TES: Skyrim, Battlefield V and many other games on the Ryzen 7000 integrated graphics

You can play TES: Skyrim, Battlefield V and many other games on the Ryzen 7000 integrated graphics

The first reviews of Ryzen 7000 processors that began to be published yesterday weakly touched on the built-in RDNA 2 graphics, which is present in all AMD next generation chips.
The manufacturer itself has previously confirmed that the Ryzen 7000 \"embedded\" is not intended for gaming, but only for basic tasks.
HotHardware and TechPowerUp portals have looked into whether this is really the case.
It turns out that you can play games on the \"embedded\" of the new Ryzen.Image source: HotHardwareFirst, let us remind you that the embedded graphics is part of the chiplet with output interfaces.
It is made according to the 6-nm process.
The basis of the graphics core is a pair of RDNA 2 blocks, which together have 128 stream processors.
The clock frequency is 2200 MHz.
According to journalists, the integrated graphics in Ryzen 7000 processors is really not suitable for playing modern games.
Especially when it comes to gaming titles that rely heavily on graphics performance.
They refuse to run at an acceptable level of performance even when using minimum quality settings.
For example, Middle Earth: Shadow of War, as well as Watch Dogs: Legion and Call of Duty: Modern Warfare II.Raster performance \"embedded\" processors AMD new generation is at the level of graphics cards Radeon HD 4850, Radeon HD 5670 or Radeon HD 7730 and not too far behind the Radeon R7 250.
These are very old video cards released more than 10 years ago.
The GeForce MX150 and Intel UHD 750 are also cited as examples with comparable performance.
In the first case, the performance is almost the same, in the second, AMD graphics are \"slightly faster\".
But you shouldn't expect more from two RDNA 2 computational units, working at 2200 MHz.
At the same time, new AMD Ryzen 7000 integrated graphics handles old and not too demanding games.
Battlefield V, Unreal Tournament, Counter-Strike: Global Offensive, Starcraft II, The Elder Scrolls (including Skyrim) and thousands of other classic gaming titles should run on \"embedded\" without problems.
Especially in reduced to 720p resolution.
It is also noted that the graphics cope well with platformers, including modern.Running more or less modern games on the built-in graphics Ryzen 7000 processors is also possible.
For example, HotHardware tested it in games such as Minecraft, Neon White, Prodeus and multiplayer Risk of Rain 2.
In all of them, the Ryzen 7000 \"embedded\" showed a \"normal level of performance.\" Fortnite was not left out either.
We ran it at 1080p resolution with DirectX 11 on medium settings.
Under these conditions, the game averaged 30 frames per second.
Reducing the resolution to 1600 × 900 pixels increased the average FPS score to 50, and by lowering the settings even more, the game was able to achieve more than 60 frames per second.
Despite this, journalists would not recommend the Ryzen 7000 chip's integrated graphics as a permanent gaming solution.

Mobile 14-core Core i9-13900HK and Core i7-13700H made a mark in Geekbench - flagship showed 5.3 GHz

Mobile 14-core Core i9-13900HK and Core i7-13700H made a mark in Geekbench - flagship showed 5.3 GHz

Mobile 14-core Core i9-13900HK and Core i7-13700H made a mark in Geekbench - flagship showed 5.3 GHz

Mobile Core i9-13900HK and Core i7-13700H processor test results have appeared in Geekbench synthetic test database.
According to the test, both chips are 14-core (six powerful and eight power-efficient cores) and support 20 virtual threads.
Chips were tested in some Samsung laptops.Image source: WccftechAccording to the test details, the Core i9-13900HK is able to automatically overclock to 5.3GHz.
The Core i7-13700H was able to clock up to 4.95GHz.
Recall that the current mobile Core i9-12900HK and Core i7-12700H generation Alder Lake have maximum clock speeds of 5.0 and 4.7 GHz respectively.
Interestingly, the novelties are claimed to have 1.25 MB of L2 cache per core, but the Raptor Cove cores architecture implies 2 MB of L2 cache per core.
Image source: GeekbankshThe performance of the new processors, especially the multi-threaded ones, is not impressive yet.
In the Geekbench database, you can find higher results for the same Core i9-12900HK.
However, the new chips are engineering samples, so it's too early to definitively judge their performance.Core i9-13900HK results (top) and Core i7-13700H (bottom) Single-threaded processor performance comparison.
Image source: WccftechIntel has previously confirmed that the mobile Raptor Lake series processors will be released by the end of this year.
However, the company did not specify whether it applies to all models of the new series.
According to VideoCardz, the first notebooks based both on the new Intel mobile processors, and AMD next-generation processors are expected to be announced at CES 2023 early next year, along with the announcement of mobile video cards NVIDIA Ada Lovelace and AMD RDNA 3.
Multi-threaded processor performance comparison.
Image source: Wccftech

9.21.2022

According to Geekbench, the Ryzen 9 7900X is a third faster than its predecessor

According to Geekbench, the Ryzen 9 7900X is a third faster than its predecessor

According to Geekbench, the Ryzen 9 7900X is a third faster than its predecessor

Performance data for the AMD Ryzen 9 7900X 12-core processor has appeared in the Geekbench 5 synthetic test database.
It remained the only chip in the new Ryzen 7000 line of processors for which no independent performance tests were published in this benchmark.
Earlier leaks confirmed the performance of the six-core Ryzen 5 7600X, the eight-core Ryzen 7 7700X and the flagship 16-core Ryzen 9 7950X.Image source: WccftechThe Ryzen 9 7900X has a base frequency of 4.7 GHz.
On one of the eight cores, it can automatically overclock to 5.6 GHz.
The chip has a claimed TDP of 170W, 76 MB of cache memory (64 MB of L3 and 12 MB of L2) and supports 24 virtual threads.
Image source: GeekBenchThe processor was tested with Gigabyte X670E Aorus Master motherboard equipped with 32 GB of DDR5 memory.
The chip scored 2167 points in single-core test and 18,622 points in multi-core test.
Novelty on Zen 4 architecture was 30% faster in single threaded and 33% faster in multi-threaded tests than its predecessor, Ryzen 9 5900X.
In comparison with Core i9-12900K new AMD chip was also faster.
Intel's still current 16-core flagship scored 1,901 and 17,272 points in the same tests, respectively.
Compared to engineering samples Core i7-13700K and Core i9-13900, Ryzen 9 7900X processor is faster in single threaded performance, but inferior in multi-threaded performance.Comparison table for single threaded performanceAmd Ryzen 7000 processors and motherboards for new Socket AM5 platform on senior AMD X670E and X670 chipsets will go on sale September 27.

9.20.2022

The 35 watt Core i7-13700T has appeared in the Geekbench test database, and it is higher than the Ryzen 7 5800X

The 35 watt Core i7-13700T has appeared in the Geekbench test database, and it is higher than the Ryzen 7 5800X

The 35 watt Core i7-13700T has appeared in the Geekbench test database, and it is higher than the Ryzen 7 5800X

Only K-series models with unlocked multiplier are expected to be announced as part of the first wave of Intel Raptor Lake series processors.
However, a model from the future energy efficient Core T series has made its mark in the Geekbench synthetic test database.
We're talking about the 16-core and 24-thread Core i7-13700T with an estimated heat dissipation of 35 watts.Image source: VideoCardzBase frequency of the processor is 1.4 GHz.
In auto mode, it can overclock up to 4.78 MHz.
Recall that the Intel T-series processors are not designed for gaming.
It is designed for use in compact energy-efficient and embedded systems, often equipped with passive cooling systems.Image source: GeekbanenchIn single-core performance tests, Core i7-13700T showed results from 1893 to 1939 points.
In multi-core performance tests, the chip scored from 8,925 to 11,564 points.Source image: GeekbanxThe Core i7-13700T is faster than the eight-core Ryzen 7 5800X on the Zen 3 architecture.
The maximum TPD of the Raptor Lake chip is unknown, but it should be lower than that of the Ryzen model (105W nominal and 142W maximum).
Intel chip was tested with DDR4-3200 RAM, hence it did not take advantage of faster DDR5 memory support.Compared to the 10-core Core i5-12600K Alder Lake series, the Core i7-13700T showed higher single-core and almost the same multi-core performance.
Image source: VideoCardzIntel usually releases energy-efficient models of its processors after the release of the higher-end K-series models with TDP 125W and mass models without overclocking and with TDP 65W.

9.10.2022

Intel Arc A580 graphics card showed very strange results in Ashes of the Singularity gaming tests

Intel Arc A580 graphics card showed very strange results in Ashes of the Singularity gaming tests

Intel Arc A580 graphics card showed very strange results in Ashes of the Singularity gaming tests

Fresh performance data of Intel Arc A580 desktop graphics card appeared in Ashes of the Singularity gaming test database.
It's not the first time this model has been noted there.
Last month the card was tested with graphic API Vulkan.
In a fresh test, the novelty was tested using DirectX 11 and DirectX 12 graphics APIs.
The results turned out to be ambiguous.Image source: IntelNow Intel has disclosed the technical specifications of the Arc A580.
The gas pedal is built on the ACM-G10 graphics chip, which includes 24 Xe cores (3072 FP32 blocks), 24 ray tracing blocks, and 384 XMX matrix engines.
The declared GPU frequency is 1700 MHz.
It has 8 GB of GDDR6 memory with 16 GHz effective bandwidth, 256-bit bus bandwidth and 512 GB/s bandwidth.
According to Intel, Arc Alchemist graphics cards are best optimized for DirectX 12 games, with Vulkan APIs offering slightly worse performance.
Performance in games based on old DirectX 11 will be the worst.
It is interesting that in recent gaming test Ashes of the Singularity our Arc A580 showed mediocre performance not only in DirectX 11 mode, but also with DirectX 12 API.
Its performance was significantly worse than when using Vulkan API.
In all cases the game test was run at 1080p resolution with average image quality settings.
When using Vulkan, the card scored 9,300 points with an average frame rate of 95 FPS.
Intel Arc A580 in Vulkin API modeThe card scored 2,800 points in DirectX 11 mode with an average frame rate of 30.6 FPS.
In DirectX 12 mode, the card scored 6,600 points and 68.6 FPS.
It is interesting that in DX11/12 mode it was worse than entry level Arc A380.
Intel Arc A580 in DirectX 11 API modeIntel Arc A580 in DirectX12 modeThe below is a performance comparison table with other graphics cards provided by Tom's Hardware.There may be several reasons for such strange results.
Maybe the CPU in the system used to test the Arc A580 is so severely limiting the performance that the GPU, which should be three times faster than the Arc A380, ends up being slower.
Another reason could be the lack of optimization of its graphics driver or the disabled Resizable BAR (ReBAR) feature, without which the performance of Arc video cards drops significantly.
Another assumption is that the Arc A580 tested, which for some reason has an RI suffix in its name, is a mobile version of a power-limited graphics card.
It's also possible that it's not a full-fledged desktop Arc A580, and someone tricked the test into mistaking another graphics card for a desktop model.

9.09.2022

The eight-core Ryzen 77700X was faster than the 16-core Core i9-12900K in the CPU-Z test

The eight-core Ryzen 77700X was faster than the 16-core Core i9-12900K in the CPU-Z test

The eight-core Ryzen 77700X was faster than the 16-core Core i9-12900K in the CPU-Z test

A database of the CPU-Z utility's built-in performance test revealed the performance of the eight-core AMD Ryzen 7 7700X processor on the 5nm Zen 4 architecture.
Soon the record was deleted, but vigilant users managed to take screenshots.Image source: AMDAll eight cores of AMD processor in the test worked at the same frequency of 5425 MHz.
This is 25 MHz higher than the claimed auto overclocking frequency for this model.
The TDP rating of the Ryzen 7 7700X processor is 105W.
It supports AVX512IF instructions, DDR5 memory and PCIe 5.0 interface.
AMD Ryzen 7 7700X specifications.
Image source: ValidX86 / TUM_APISAKThe processor was tested on Gigabyte X670E AORUS Master motherboard, and with DDR5-6400 memory with CL30 timings.
With this configuration, the chip scored 774 points in CPU-Z single-core performance test, and 8381 points in multi-threaded test.
Ryzen 77700X single-core performance test with CPU-Z.
Source image: ValidX86 / TUM_APISAKWith these single-core performance results, the novelty is 21% faster than the eight-core Ryzen 7 5800X and 20% faster than the 16-core Ryzen 9 5950X on Zen 3 architecture.
In addition, the Ryzen 7 7700X was 1% faster than the 10-core Core i5-12600K, but inferior to the 12-core Core i7-12700K.
Ryzen 7 7700X multicore performance test by CPU-Z.
Image source: ValidX86 / TUM_APISAKMultithreaded performance of the Ryzen 7 7700X in this test was 1 percent better than the 16-core Core i9-12900K and 8 percent better than the Core i7-12700K.
Moreover, new AMD Raphael CPU is 28% faster than Ryzen 7 5800X.
AMD new product also showed up in Geekbench synthetic test database.
However, in this case we used a different motherboard (ASUS ROG Crosshair X670E Hero) and a different RAM configuration (64 GB DDR5-6000 memory).
Source image: GeekbenchWith results of 2209 and 14,459 points in single-threaded and multi-threaded performance tests, the novelty in the first case was 16% faster than the Core i7-12700K model, and in the second - 2% faster.Ryzen 7 7700X processor along with the rest of the Ryzen 7000 series goes on sale September 27.
Reviews of the new generation chips will start to appear one day earlier.

9.05.2022

Intel Core i9-13900 embedded graphics were slower than AMD Radeon Vega 10 in Geekbench OpenCL test

Intel Core i9-13900 embedded graphics were slower than AMD Radeon Vega 10 in Geekbench OpenCL test

Intel Core i9-13900 embedded graphics were slower than AMD Radeon Vega 10 in Geekbench OpenCL test

The Geekbench OpenCL synthetic test database has revealed performance data on Intel Core 13th generation processor embedded graphics.
The test was conducted on a certain HP PC equipped with an Intel Core i9-13900 processor.
In the OpenCL graphics test, Intel's Embedded showed a score of 9498 points.
This is roughly 93% lower than the NVIDIA GeForce RTX 3070 discrete graphics card.image source: PC-Magazin.eThe Raptor Lake processor has 32 execution units running at 1.65 GHz.
The current flagship Core i9-12900 has an Intel UHD Graphics 770 \"embedded\" with the same number of execution units, but running at 100 MHz lower or 6.5% slower.
One would expect the Raptor Lake integrated graphics core to be about 6.5 percent slower than the Alder Lake graphics due to this.
But in practice, the iGPU of the 13th generation Intel Core processor turned out to be only 2.6% faster.
Source image: GeekbankshIn comparison with the closest competitors, Raptor Lake graphics turned out to be slower than AMD Radeon Vega 10, NVIDIA GeForce MX350 entry-level discrete mobile graphics card or the same GeForce GTX 750 and Radeon RX 540 \"embedded\".
On the other hand, keep in mind that the test was most likely conducted on the Raptor Lake engineering sample and the final figures of the integrated graphics performance may differ, though hardly significantly.Image source: Tom's HardwareAdd that the test once again confirmed that the Core i9-13900 processor has 24 physical cores with support for 32 virtual threads and runs in the 2 to 5.3 GHz frequency range.
It is also reported to carry 36 Mbytes of L3 cache memory on board.The official announcement of the Intel Raptor Lake series of chips will take place at the end of this month.
They are rumored to go on sale in mid-October, along with new Intel 700-series motherboards.

Intel's own tests show the Arc A770 performs better at ray tracing than the GeForce RTX 3060

Intel's own tests show the Arc A770 performs better at ray tracing than the GeForce RTX 3060

Intel's own tests show the Arc A770 performs better at ray tracing than the GeForce RTX 3060

To prove its claim that its Arc Alchemist graphics performs better at ray tracing than comparable NVIDIA GeForce RTX graphics cards, Intel has revealed performance of its flagship Arc A770 gas pedal compared to a GeForce RTX 3060 graphics card.
Image source: IntelAccording to Intel's internal tests, the Arc A770 is on average 14 percent faster than the GeForce RTX 3060 in games at 1080p resolution and ray tracing.
However, in at least four games from the list of tested titles NVIDIA's solution was still faster.
Intel Arc A770 vs NVIDIA GeForce RTX 3060The GeForce RTX 3060 graphics card has 28 RT cores.
Intel has not yet announced official specs for the Arc A770.
However, it is assumed that it uses a full-fledged version of the processor ACM-G10 with 32 cores Xe and the same number of blocks of hardware ray tracing.Image source: VideoCardzIntel also first shared the official performance test cards Arc A-series with active technology time scaling image Xe Super Sampling (XeSS).
It should be recalled that the company had promised to release it \"in early summer.
The summer ended and the fall came, but so far the support for XeSS has not appeared in any game.
Today, the company published the data on the performance of XeSS, which operates in two modes: Balanced and Performance.
In total there will be four modes.
The other two, Quality and Ultra Quality, were not included in the comparison.
The \"Performance\" mode allows in some cases (Hitman 3 and Ghostwire Tokyo) a doubling of frames per second.
In general, however, the FPS gain range is between 49% and 113% over the original resolution.
In \"Balance\" mode, the frame rate increases from 26% to 77%.
The Intel Arc Alchemist architecture uses a new type of Thread Sorting Unit cores, which are responsible for asynchronous ray tracing.
In other words, they help maximize the use of Xe cores by distributing tracing tasks in the graphics pipeline to improve efficiency and performance.Arc Alchemist arc-ray-tracing-hardware-1 ray tracing pipeline on Intel Arc Alchemist graphics cards.
jpg See all images (4) asynchronous-ray-tracing-1.jpg INTEL-ARC-GPU-RAY-TRACING-1.jpg thread-sorting-ray-tracing-1.jpg See all images (4) The company has also added that support for its ray tracing technology will appear in the game Gotham Knights, which will release on October 21.
I'd like to believe that by that time Intel's senior graphics cards will be available for sale.

8.31.2022

Ryzen 9 7950X results hit the Geekbench database, and they are worse than the engineering Core i9-13900K

Ryzen 9 7950X results hit the Geekbench database, and they are worse than the engineering Core i9-13900K

Ryzen 9 7950X results hit the Geekbench database, and they are worse than the engineering Core i9-13900K

Geekbench synthetic test database has posted performance test results for the flagship 16-core Ryzen 9 7950X processor.
Previously reported performance in the same benchmark of the junior six-core Ryzen 5 7600X model.Image source: AMDRyzen 9 7950X was tested on ASUS ROG Crosshair X670E Extreme motherboard paired with 32GB of DDR5-6000 RAM.
As part of the test, the processor automatically overclocked to 5759 MHz, just above the manufacturer's claimed 5.7 GHz.
Ryzen 9 7950X has 16 physical cores with support for 32 virtual threads.
The same number of cores have still current flagship processors AMD Ryzen 9 5950X and Intel Core i9-12900K.
At the same time the new AMD chip has fewer cores than the future flagship Core i9-13900K processor.
The latter will be equipped with 24 physical cores, but 16 of them are energy efficient and do not support Hyper-Threading.Ryzen 9 7950X specifications.
Source: Geekbanksh In the single-core performance test, the Ryzen 9 7950X scored 2,217 points.
This is just below the result announced by AMD itself (pictured below).
Multi-core performance index of the chip is 24,396 points.AMD declared Ryzen 7000 processor performanceMost models of future Intel Core 13th generation processors have already been tested in this benchmark (though, as engineering and qualification samples), the results of new AMD product can be compared with them.Ryzen 9 7950X performance.
Source: GeekbenchThe most recent entry in the Geekbench 5 database shows single-core performance of the Intel Core i9-13900K Raptor Lake series is 4.4 percent better.
The multi-core performance of the Intel chip is 8.8 percent higher than the Ryzen 9 7950X.
At the same time, the new AMD processor is 31 and 47.8% faster in single-threaded and multi-threaded tasks than the Ryzen 9 5950X, and 14.5% and 41.2% faster in the same tests than the Core i9-12900K.Image source: VideoCardzIt should be recalled again that only engineering and qualification samples of 13th generation Intel Core processors took part in Geekbench tests.
This means that their final performance figure may differ from the performance of the off-the-shelf consumer variants.

8.29.2022

Chinese server processors KaiSheng KH-40000 were noticeably slower than modern desktop Intel and AMD

Chinese server processors KaiSheng KH-40000 were noticeably slower than modern desktop Intel and AMD

Chinese server processors KaiSheng KH-40000 were noticeably slower than modern desktop Intel and AMD

The performance data of KaiSheng KH-40000 processors from Chinese company Zhaoxin appeared in Geekbench 5 synthetic test database.
Formally, this series of chips has not yet been presented.
It will include models with up to 16 cores.
As the first tests show, the performance of KaiSheng KH-40000 has increased against its predecessors, but it is not as good as modern AMD and Intel desktop processors.Image source: ZhaoxinThe Zhaoxin company, owned by Via Technologies and the Shanghai government, has continued to use Via microarchitectures in its products (or more specifically, Centaur Technology, which was absorbed by Via) since the middle of the last decade.
The upcoming KaiSheng KH-40000 series of data center processors use the CentaurHauls microarchitecture, which some technology experts say resembles the Intel Haswell architecture from 2013.The 16-core KaiSheng KH-40000/16 and 12-core KaiSheng KH-40000/12 were noted in Geekbench 5 tests.
Both chips run at a frequency of 2.20 GHz.
It is known that the senior model has 32 MB of L3 cache memory, the junior model has 24 MB of L3 cache memory.
The senior model supports SMT technology, so the processor has 32 virtual threads.
Based on the data collected by the Geekbench 5 test, the Zhaoxin KaiSheng KH-40000/16 and KaiSheng KH-40000/12 models are very similar to the never-released Centaur CHA processor reported earlier.
Nevertheless, there is a difference between the chips.
The difference is half the number of cores of the Centaur CNS, its lack of SMT support, and the use of TSMC's more mature N16 manufacturing process.
The new KaiSheng KH-40000 chips are believed to use the more modern TSMC N7 process technology.Although the new KaiSheng KH-40000 processors on the CentaurHaul architecture operate at 2.20 GHz, in tasks related to single-threaded performance, they prove to be significantly faster than the previous generation chips on the LuJiazui architecture, which operate at a higher frequency of 2.70 GHz.
This statement is true both for integer operations, where the new chips are 22% faster, and in floating point operations, where their performance is 75% better.On the background of predecessors, which in the comparison test (results below in the table) represents the eight-core Zhaoxin KX-U6780A model, KaiSheng KH-40000 processors show themselves significantly better.
However, in comparison with any modern AMD and Intel processors, they lose out cleanly.
For example, their single-threaded performance is even lower than AMD's Bulldozer/Piledriver chips from 2012.
There is also a very significant difference in multi-threaded performance between the KaiSheng KH-40000 12-core and 16-core models.
In theory, a chip with 16 cores with support for 32 threads can indeed handle 2.66 times as many threads as a processor that has 12 cores with 12 threads.
However, as the results above show, the KaiSheng KH-40000/16 is even faster: 2.69 times in Integer tasks and 2.84 times in Float operations.
The senior processor has only four more physical cores than the junior one, but its performance is almost three times better.It should be noted that Windows 10 and Windows 11 operating systems do not always work optimally with the task scheduler when an unfamiliar multicore processor is involved.
Therefore, the results of the 12-core KaiSheng KH-40000/12 model, which was tested in Windows 10 Pro, may not reflect the real potential of the processor against the older model, which was tested in a more suitable environment.
Even so, the multi-core performance of the 12-core model on the CentaurHoals architecture was expectedly higher than that of the chip on the preceding LuJiazui architecture.
The latter chip not only has fewer cores, but also uses an older process technology and has a smaller L3 cache.
In Integer and Float operations the difference in favor of the newcomer was 40 and 78%, respectively.
However, as mentioned above, the overall performance level of KaiSheng KH-40000 processors and their predecessor Centaur CHA by modern standards is very modest.
For example, the level of multi-threaded performance of 12-core KaiSheng KH-40000/12 in Windows environment is comparable to the eight-core AMD FX-8350.

8.26.2022

Ryzen 7 7700X engineering sample was up to 33% faster than Ryzen 7 5800X in Cinebench R20 test

Ryzen 7 7700X engineering sample was up to 33% faster than Ryzen 7 5800X in Cinebench R20 test

Ryzen 7 7700X engineering sample was up to 33% faster than Ryzen 7 5800X in Cinebench R20 test

Chinese tech blogger Extreme Player has shared performance data about AMD Ryzen 7 7700X eight-core processor engineering sample, demonstrated in Cinebench R20 synthetic test.
The chip was up to 33% faster than the Ryzen 7 5800X model on the still current Zen 3 architecture.Image source: ExtremePlayerIn Cinebench R20 engineering sample Ryzen 7 7700X, running at 4.5 GHz, scored 773 and 7701 points respectively in single-threaded and multi-threaded tests.
In the former, this is 23% more than the eight-core Ryzen 7 5800X, and in the latter, 33% more.
In single-threaded performance engineering sample Ryzen 7 7700X surpassed even the current flagship Ryzen 9 5950X - the difference is 19% in favor of news.From earlier leaks we know that the engineering sample of 16-core Intel Core i7-13700K future series processors Raptor Lake scored 814 points in single-threaded test Cinebench R20.
In turn, multi-threaded performance of the upcoming Intel novelty is estimated to score 11 243 points in the same test.
This is 46% higher than the result of Ryzen 7 7700X engineering sample.
Source image: VideoCardzThe official Ryzen 7000 processors will be announced in a few days, on August 29.

6.22.2022

Intel Arc A380 fails miserably in first review - up to 26% slower than Radeon RX 6400

Intel Arc A380 fails miserably in first review - up to 26% slower than Radeon RX 6400

Intel Arc A380 fails miserably in first review - up to 26% slower than Radeon RX 6400

Intel has lifted the ban on publishing reviews of its Arc A380 desktop graphics card.
And in the first independent test the novelty with a bang lost even the entry level model AMD Radeon RX 6400.Image source: Bilibili / ShenmedounengceThe non-standard Gunnir Arc A380 Photon graphics card was tested by Chinese reviewer Shenmedounengce, who checked the novelty in synthetic 3DMark tests, as well as in a number of games.
And if in the first case, the novelty showed itself quite well, then with games the situation was much worse.
The card has been tested with Core i5-12400 processor, Intel B660 motherboard and DDR4 RAM.
For comparison with the Intel Arc A380, the reviewer used GeForce RTX 3050, GeForce GTX 1650, as well as Radeon RX 6500 XT and RX 6400 graphics cards.
The Intel Arc A380 is based on the ACM-G11 graphics processor, which includes eight Xe cores running at 2.45 GHz.
The gas pedal got 6 GB of GDDR6 memory with 92 bit bus support and 15.5 Gbit/s per pin.In 3DMark benchmarks Intel's card performed well.
In the Timespy test, Intel's new product confidently beat the GeForce GTX 1650, Radeon RX 6400 and RX 6500 XT graphics cards, and not too significantly behind the GeForce RTX 3050.
In the 3DMark Port Royal test, which evaluates gas pedals in ray tracing, the Intel Arc A380 was almost twice as fast as the Radeon RX 6400 and RX 6500 XT, but still was more than 3.5 times slower than the GeForce RTX 3050.
The game tests used were League of Legends, PUBG, GTA V, Shadow of Tomb Raider, Forza Horizon 5 and Red Dead Redemption 2.
All were run at 1080p resolution.
Unfortunately, in none of them the Arc A380 was able to catch up even with the youngest Radeon RX 6400 model.
At best, the gap was about 7%, and in some games Intel gas pedal lost in terms of performance by 26%.
Recall that the launch of the Intel Arc A380 took place only in China.
In the official sale of the new product has not yet appeared.
Its declared recommended price is about $150.

6.08.2022

Intel Arc A730M mobile graphics card is barely faster in games than GeForce RTX 3050

Intel Arc A730M mobile graphics card is barely faster in games than GeForce RTX 3050

Intel Arc A730M mobile graphics card is barely faster in games than GeForce RTX 3050

Despite the good performance of Intel Arc A730M mobile graphics card in 3DMark Time Spy synthetic test, where it performed even better than GeForce RTX 3070, in real games the Intel gas pedal was barely faster than mobile GeForce RTX 3050.
One of the Chinese insiders has shared information about the new product performance.Image source: VideoCardzWeibo user with a nickname Golden Pig Upgrade has tested Intel Arc A730M in a number of games.
These included Assassin's Creed: Odyssey, Metro Exodus and F1 2020.
The tests were conducted in two resolutions.
Unfortunately, the source did not specify whether the tests used Intel DTT (Dynamic Tuning Technology), which dynamically distributes power between the laptop's CPU and graphics processors.
As we found out earlier, when testing the mobile Intel Arc A350M, this technology negatively affects performance, which can be attributed to a software flaw on the part of Intel.
Intel Arc A730M in Metro Exodus at 1080p resolution.
Source: Weibo Intel Arc A730M in Metro Exodus at 1440p.
Source: WeboIn Metro Exodus, the Intel Arc A730M averaged 70fps at 1080p and 55fps at 1440p.
High graphics settings were used in both cases.
Such performance is higher than the GeForce RTX 2070, which shows 66 frames per second at 1080p resolution, but worse than the GeForce RTX 3060.
The latter averages 80 frames per second at the same resolution and graphics settings.
Intel Arc A730M in F1 2020 at 1080p resolution.
Source: Weibo Intel Arc A730M in F1 2020 at 1440p.
Source: WeiboIn F1 2020, the Intel card showed 123 frames per second at 1080p and 95 frames per second at 1440p.
By comparison, the GeForce RTX 3050 averages 120 frames per second at 1080p resolution at high graphics settings.
Intel Arc A730M in Assassin's Creed: Odyssey at 1080p.
Source: Weibo Intel Arc A730M in Assassin's Creed: Odyssey at 1440p.
Source: WeboIn Assassin's Creed: Odyssey, the Intel gas pedal averaged 38 (1080p) and 32 (1440p) frames per second.
The source also notes that during the tests, almost all games saw short-term significant drops in the gas pedal performance to below 10 frames per second, which may indicate poor software optimization.According to Intel Arc Control Center, the Arc A730M graphics chip was running at 2050 MHz and consumed 92 watts of power.
We should add, however, that these figures were observed when running the Furmark test, which doesn't exactly reflect the performance of video cards in real-world conditions.
Intel Arc A730M.
Intel Arc Control Center data.
Source: WeboSo far, everything points to the Arc A730M's performance level being next to the GeForce RTX 3050 and GeForce RTX 3060.
The non-optimized gaming performance of Arc graphics cards, as this is not the first such test shows, may be the main reason why Intel is in no hurry with the global release of its discrete graphics cards.