According to reports, the driver released by AMD last month is optimized for the power consumption of the RX 7900 series when playing video. Before optimization, the RX 7900 XT consumes 71W to play YouTube videos, and the RX 7900 XTX GPU consumes 81W. After optimization, the power consumption of the two cards is reduced to 46W and 54W respectively, which is still not as good as the RTX 4080 (34W) or the Radeon RX 6900 XT (30W).

In addition, foreign media ComputerBase found that the game power consumption of the RX 7900 under the new driver has also been optimized, but only in scenes where the frame rate is limited.

As shown above, in the Doom Eternal game at 1440p resolution, the RX 7900 XTX is 57% more efficient (down 90W) and the RX 7900 XT is 34% more efficient (down 57W) with the new driver when the frame rate is limited to 144fps. When the frame rate is not limited, there is no significant change in power consumption between the two cards under the old and new drivers.

The post AMD optimizes the power consumption of Radeon RX 7900 series graphics cards for gaming appeared first on TechGoing.

The Biostar RX 7900 XT graphics card has a frequency of 2395MHz. It adopts a simple three-fan design. From the side, the thickness is more than 3 slots. It is powered by dual 8pin interfaces, equipped with three DP 2.1 interfaces and one HDMI interface.

The Biostar RX 7900 XTX graphics card is a public version design, with a GPU frequency of 2500MHz, dual 8pin power supply, two DP 2.1 interfaces and one HDMI interface, and the thickness is only two slots, thicker than the RX 7900 XT. Smaller.

AMD radeon RX 7900 XT / XTX two flagship graphics cards have been officially launched. The first batch includes non-public models of ASRock, ASUS, Gigabyte, Sapphire, and XFX. The prices start from 7399 CNY and 7999 CNY respectively. The price of the product has not yet been announced.

The post Biostar Releases AMD Radeon RX 7900 Series Graphics Cards with Reference Design appeared first on TechGoing.

It is said that there will be more than 30,000 public versions on the shelves around the world on the day of the sale, so the rest of the GPUs are likely to be sold to AIB manufacturers, such as Sapphire, XFX, Xunxun, ASRock, etc., which have been warming up these days. It is expected that major manufacturers will bring a large number of non-public graphics cards, which are expected to start selling in mid-December.

The AMD Radeon RX 7900 XTX and Radeon RX 7900 XT graphics cards are based on the Navi 31 GPU of the RDNA 3 architecture, with a single GCD of 48 WGP, 96 computing units, 12 SA and 6 SE, and the two graphics cards will be officially launched tomorrow. The prices are 7999 CNY and 7399 CNY respectively.

From the statistical data, Nvidia’s total shipments as of mid-November are about 160,000 GPUs, including 130,000 RTX 4090 (AD102) and 30,000 RTX 4080 (AD103) chips. It is estimated that this number has now exceeded 200,000, but In any case, the supply of RTX 4080 is significantly less than that of RTX 4090, after all, RTX 4090 is more expensive.

"Here are the parameters of the RX 7900 series graphics cards:

The flagship model RX 7900 XTX has 96CU, 6144 stream processors, GPU frequency 1900-2500MHz, single-precision performance up to 61TFLOPS. The video memory is 24GB 20Gbps GDDR6, the bandwidth can reach 960GB/s, plus 96MB unlimited cache, the equivalent bandwidth can reach 3.5TB/s. The TBP of the graphics card is 355W, with power requirements starting at 800W. The RX 7900 XTX reference version has a length of 287 mm, a 2.5 slot thickness, and interfaces including DP 2.1, HDMI 2.1 and USB-C.

The sub-flagship model RX 7900 XT has 84CU, 5376 stream processors, GPU frequency of 1500-2400MHz, and single-precision performance of up to 52TFLOPS. The video memory is 20GB 20Gbps GDDR6, the bandwidth can reach 800GB/s, plus 80MB unlimited cache, the equivalent bandwidth can reach 2.9TB/s. The TBP of the graphics card is 300W, with power requirements starting at 750W. The length of the RX 7900 XT reference version is 276 mm, the thickness of the 2.5 slot, and the interface includes DP 2.1, HDMI 2.1 and USB-C."

The post AMD Q4 shipped 200,000 RDNA 3 Navi 31 GPUs for RX 7900 XT / XTX graphics cards appeared first on TechGoing.

Judging from the renderings, the graphics card has a complete non-princess board design with three 8-pin power connectors. This is a common trait of all the premium RX 7900 XTX models we’ve seen so far. In addition, there is a dual BIOS switch, but there is no complete specification information on the graphics card at present.

Gigabyte is still preparing a Gaming OC graphics card. The graphics card is relatively affordable in terms of price, 2.5 slots thick, with a triple-fan heatsink and dual 8-pin power connectors. Interestingly, this model will also get a dual BIOS feature. Gigabyte will launch the Gaming OC series of RX 7900 XTX and RX 7900 XT graphics cards.

The post Gigabyte expected to launch Radeon RX 7900 Series AORUS Elite and Gaming OC Graphics soon appeared first on TechGoing.

The first GeekBench benchmark scores for the RX 7900 XTX public graphics card have been revealed, and surprisingly, its benchmark score is far lower than AMD’s published numbers. The GeekBench running score of the RX 7900 XT public graphics card was also exposed today, and the performance was also lower than expected.

GeekBench running scores show that the RX 7900 XT scored 191822 points in OpenGL, which is 16% slower than the RX 7900 XTX. It has 5376 SPs or 84 Compute Units (CUs). That’s 12.5 percent less than the compute unit of the flagship XTX. With such a difference in performance, more users will choose XTX when the price difference is only 600 yuan.

The flagship model RX 7900 XTX has 96CU, 6144 stream processors, GPU frequency 1900-2500MHz, single-precision performance up to 61TFLOPS. The video memory is 24GB 20Gbps GDDR6, the bandwidth can reach 960GB/s, plus 96MB unlimited cache, the equivalent bandwidth can reach 3.5TB/s. The TBP of the graphics card is 355W, with power requirements starting at 800W. The RX 7900 XTX reference version has a length of 287 mm, a 2.5 slot thickness, and interfaces including DP 2.1, HDMI 2.1 and USB-C.

The post AMD Radeon RX 7900 XT public graphics card running points exposure appeared first on TechGoing.

According to foreign media, according to its sources, the AMD Radeon RX 7900 XT graphics card will not be the top model. The AMD Radeon RX 7900 XT was originally designed with 24 GB of video memory, which has now been downgraded to 20 GB. It is expected that the RX 7950 XT model with a full Navi 31 GPU + 24GB of video memory will benchmark against Nvidia’s RTX 4090 Ti model in the future.

According to previous revelations, the AMD Navi31 GPU will use a GCD + four MCD chip design, with 12288 stream processors and support for 384bit bit-width video memory.

AMD has now announced that it will hold the RX 7000 series graphics card conference at 4:00 p.m. Eastern Time on November 3. The RX 7000 series graphics cards will have the following highlights:

5nm process node

Advanced Chiplet Packaging

New Architecture Compute Unit

Optimized graphics pipeline

Next Generation AMD Infinite Cache

50% improvement in performance per watt compared to RDNA 2

The post AMD’s Radeon RX to have more memory: 20GB for the 7900 XT and 24GB for the 7950 XT appeared first on TechGoing.

According to the data, the AMD Radeon RX 7900 XT graphics card can deliver more than 7 TB per second of bandwidth. While the discussion is classified as speculative or theoretical, in this case, it is no less than a fascinating insight into AMD’s new graphics card.

In the post, the Bright Iron user does his calculations using the information found in AMD’s RDNA 3 Navi 31 flagship GPU, which offers 6 MCDs and up to 384 MB of Infinity Cache in a 3D V-Cache package.

His theory is based on information already available, such as the AMD Radeon RX 6950 XT, which is the best performer in the company’s current GPU lineup, offering 16 GB of GDDR6 memory with the help of a graphics card clocked at 18 Gbps 256-bit bus. This combination of technologies results in a bandwidth speed of 576 GB/s.

Bright Iron goes on to discuss bandwidth dwarfed by NVIDIA’s GeForce RTX 3090 GPU, which offers 936 GB/s, compared to 1,008 GB/s for the company’s RTX 3090 Ti.

Bright Iron went on to explain that AMD’s bandwidth performance is quite impressive compared to NVIDIA because of the company’s use of Infinity Cache technology. Infinity Cache is a memory that is separate from the main memory located on the AMD GPU. It is typically small in size compared to the main memory on the graphics card, but due to its proximity to the chip, Infinity Cache can be pulled at a higher rate.

MD Infinity Cache provides up to 128 MB of memory capacity and 2 TB/s of memory bandwidth. The AMD Radeon RX 6850 XT offers 512 GB/s of bandwidth as standard until you add 128 MB of Infinity Cache, increasing the bandwidth by another 1,281 GB/s and providing users with a total of 1,793 GB/s of bandwidth, far more than NVIDIA can currently produce.

The newer RDNA 3 architecture will allow Infinity Cache to reach 384 MB, which is three times higher than the previous RDNA 2 architecture. The chip can generate 192 MB from the standard Cache while utilizing another 192 MB of 3D V-Cache packaging to achieve 384 MB for the entire Infinity Cache.

Bright Iron has calculated that this new amount of Infinity Cache will yield three times the bandwidth (excluding the main memory bandwidth) assuming it runs at the same speed as RDNA 2. RDNA 2’s 128MB Infinity Cache has a maximum bandwidth of 2 TB/s (terabytes per second), so we see that RDNA 3’s 384MB is 6 TB/s.

However, it could potentially run much faster, so 6 TB/s is probably the minimum requirement; again, this does not include the additional bandwidth of main memory, which is claimed to be 24GB of GDDR6 and a 384-bit bus (don’t confuse this number with 384MB of Infinity Cache).

Memory @ 16Gbps = 16,000,000,000 bps (bits per second)

16,000,000,000 bps x 384 = 6,144,000,000,000 bps

6,144,000,000,000 bps ÷ 8 bits = 768,000,000,000 Bps (Bytes per second)

768,000,000,000 Bps ÷ 1,000,000,000,000 bytes = 0.768 TB/s

0.768 TB/s + 6 TB/s = 6.768 TB/s

The post AMD Radeon RX 7900 XT theoretical bandwidth up to 6 TB/s or 7 TB/s appeared first on TechGoing.