With more and more developers starting to run ROCm on Radeon and Radeon PRO GPUs for AI and ML work, AMD launched the latest ROCm 5.6 and announced ROCm support for consumer graphics cards.

It is said that AMD will first adapt ROCm for the RX 7900 XTX and Pro W7900 RDNA 3 GPUs. The official version will be launched this fall (ROCm 5.6), and more models will be added later.

This is definitely big news for AMD users, especially those who need AI productivity. Not only that, it is noticed that ROCm 5.6 will also bring a series of new features and enhancements, especially for AI and HPC workloads, including:

 Integrate the Hugging Face unit test suite into ROCm QA

 Gradually support OpenAI Triton in PyTorch 2.0 inductor mode;

 Enable OpenXLA support for PyTorch, TensorFlow, and JAX via ROCm;

ROCm 5.6 also introduces improvements to math libraries such as FFT, BLAS, and solvers, which are the foundation of HPC applications, and enhancements to ROCm’s development and deployment tools, including installation, ROCgdb (CPU-GPU integrated debugger ), ROCm analyzer and documentation.

AMD also stated that it will continue to work on further optimizing the framework and back-end compilers for better performance, including MLIR infrastructure improvements, as the basis for AMD’s support for OpenAI Triton and OpenXLA compilers, and will continue to be added on the AMD hub of Hugging Face More open source AI models optimized for AMD solutions.

For HPC users, AMD already offers some solutions on the AMD Infinity Hub to allow customers to better build HPC application containers, and future ROCm releases will increase the number of HPC applications supported on AMD Instinct solutions.

The post AMD will bring ROCm support to RDNA 3 consumer graphics cards appeared first on TechGoing.

In the leaked images from Moore’s Law Is Dead, we can see a Navi 31 GPU in the same package seen in the Navi 32 GPU leak. The Navi 31 GCD can be seen in the middle, which has the same size of 304.35mm2, while the 6 MCDs have a chip size of 529mm2. But the main difference is that the Navi 31 GPU is housed in a more compact package measuring just 40x40mm, compared to the stock Navi 31 package which measures 55×47.5mm2. Package area is reduced by 39%.

AMD Navi 31 package size – 55×47.5mm2 = 2612.5mm2

AMD Navi 32 package size = 40×40.0mm2 = 1600.0mm2

MLID explained that AMD has sent the Navi 31 RDNA 3 GPU chip in the Navi 32 package to its partners to evaluate thermal testing before the Navi 32 is ready. The GPU also features two virtual MCD chips, meaning this Navi 31 GPU is essentially a 256-bit bus design, not a full 384-bit, or even a stripped-down 320-bit design like the 7900 XT. As such, the chip is very similar in design to the Radeon Pro W7800, but also uses the standard Navi 31 package, albeit with the two MCD chips disabled.

We can see at least two GDDR6 memory chips on the PCB, which can be used to test the ES Radeon RX 7800 XT graphics card. It is believed that the card will eventually feature at least 8 GDDR6 chips, enabling a 256-bit bus interface and 16 GB of VRAM. Now why this Navi 31 package is interesting because similar chips could be used in high-end laptop GPUs in the future. For example, we know AMD will eventually bring Navi 32 to the 7800-series GPUs, but it’s even possible that AMD could swap a Navi 32 GPU die with a Navi 31 die on the same package and offer it as a Navi 31 die. High-end Radeon RX 7900M XT for laptops with 16 GB of VRAM.

This would allow AMD to compete with NVIDIA at the high end, and possibly even the GeForce RTX 4080 eventually. It would be good to see AMD return to the high end of the market, where NVIDIA currently dominates the laptop space. MLID also said that Navi 32 is expected to launch in September and will offer the same very competitive price as the RX 7800 XT and more options.

The post AMD Navi 32 RDNA 3 GPU package image reappears Contains 6 MCDs in a compact form appeared first on TechGoing.

The AMD RX 7000 series of notebook GPUs includes the Radeon RX 7600M XT and 7600M (high-performance gaming notebooks), and the RX 7700S and RX 7600S (thin and light notebooks).

RX 7600M XT: 75~120W performance release, RDNA3 architecture, 32CU, 8GB 18Gbps GDDR6 memory

RX 7600M: 50~90W performance release, RDNA3 architecture, 28CU, 8GB 16Gbps GDDR6 memory

RX 7700S: 75~100W performance release, RDNA3 architecture, 32CU, 8GB 18Gbps GDDR6 memory

RX 7600S: 50~75W performance release, RDNA3 architecture, 28CU, 8GB 16Gbps GDDR6 video memory

In addition, AMD has launched three new RDNA2 architecture RX 6000 series notebook GPUs, namely:

RX 6550M: 50~80W performance release, RDNA2 architecture, 16CU, 4GB 18Gbps GDDR6 memory

RX 6450M: 25~50W performance release, RDNA2 architecture, 12CU, 4GB 16Gbps GDDR6 memory

RX 6550S: 35~50W performance release, RDNA2 architecture, 16CU, 4GB 16Gbps GDDR6 memory

AMD said that the RX 7600M XT notebook GPU has stronger gaming performance than the Nvidia RTX 3060 desktop version, while foreign media hothardware said that AMD representatives revealed that in some tests the performance was close to the RTX 3080 notebook GPU.

AMD showed off upcoming laptops from four vendors, including Alienware, Emdoor, Asus, and IP3, three of which are high-performance gaming laptops and one is a thin and light laptop.

The post AMD releases RX 7000 series notebook GPU: RDNA 3 architecture, up to 32CU appeared first on TechGoing.

The protagonist this time is the RDNA3 GPU, which is the RX 7000 series graphics card.

According to a new report, the RDNA3 architecture RX 7000 will not only double the performance of raster games but also improve the performance of RT light tracking more, that is, more than double. More importantly, achieving this, will not come at the cost of huge energy consumption.

Previously, AMD’s repeated emphasis on RDNA3 is that the performance per watt has increased by more than 50%.

Of course, this generation of RDNA3 will be launched after the RTX 40 series graphics cards. Let us wait and see whether it has the ability and preparation to attack.

The post AMD’s next-generation RDNA 3 graphics card conference officially announced appeared first on TechGoing.

Looking back at history, WCCFTech noted that the red team was the first to break the 1 GHz GPU clock rate – it is 28nm Tahiti GPU was used on a Radeon HD 7970 GHz Edition graphics card in 2012/GCN architecture.

Then on the Radeon RX 6000 series RDNA 2 GPU line, it hit the 3.0 GHz threshold again. Now, thanks to TSMC’s 5mn process node, AMD GPUs are apparently aiming for a new milestone of 4 GHz.

In the CPU field, the company has demonstrated a Zen 4 processor core with a significantly increased frequency with the help of TSMC’s 5nm process node. It’s no surprise to us as red teams apply their experience to other IP lineups such as RDNA GPUs.

Secondly, AMD senior vice president and technical architect Sam Naffziger emphasized that both the Radeon RX 7000 independent display and the next-generation RDNA 3 core display will usher in a series of new features, including more refined adaptive power management technology.

This setting can be specific to a specific scenario, ensuring that the GPU only consumes the power required by the corresponding workload. In addition, the updated “Infinity Cache” (Infinity Cache) provides higher density and significantly reduces memory power consumption.

The key features of the Radeon RX 7000 Series RDNA 3 GPUs are as follows.

● Use 5nm process node

● Advanced small chip packaging process

● Reserting the calculation unit (CU) architecture

● Optimized graphical pipeline

● New generation AMD unlimited cache (Infinity Cache)

● Enhanced Ray Traction performance

● Fine adaptive power management characteristics

●> 50% Perf/Watt VS RDNA 2

Finally, the line-up of AMD Radeon RX 7000 series solo displays based on the Nav 3x GPUs is expected to launch later this year – starting with the Navi 31 product line before the Navi 32 / Navi 33 SKUs.

The post Radeon RX 7000 series RDNA 3 GPU frequency is nearly 4GHz appeared first on TechGoing.