Posts for: #amd

Linux 6.6 Enhances AMD Server Performance

The Linux 6.6 kernel version seems to be delivering significant performance improvements for AMD EPYC server CPUs according to Phoronix. Tests conducted on Genoa and Genoa-X processors as well as Intel Xeon Scalable “Sapphire Rapids” processors showed that the performance of AMD EPYC servers was greatly enhanced with Linux 6.6 compared to Linux 6.5 stable. The improvements were especially notable for certain workloads.

The Linux 6.6 kernel introduces the EEVDF scheduler and workqueue enhancements that benefit chiplet-based CPUs with multiple L3 caches, such as those used by AMD. Overall, the new features in Linux 6.6 seem to be positively impacting AMD server CPU performance.

Source: Phoronix.

AMD Unveils EPYC 8004 “Siena” 4th Gen EPYC Processors

AMD has launched the AMD EPYC 8004 “Siena” processors for “intelligent edge” servers, according to Phoronix. These processors offer a lower price point while maximizing power efficiency and expanding EPYC’s deployment outside of data centers. The Siena processors have up to 64 cores / 128 threads per socket, support up to six channels of DDR5 memory, and are designed for edge computing and telco markets. They use the new “SP6” socket and have a range of TDPs from 70 Watts to 225 Watts. The EPYC 8004 series consists of P and PN parts, with PN parts being NEBS friendly. The pricing ranges from $409 to $5450 USD. The performance and power efficiency of the EPYC 8004 series are competitive against Xeon Scalable “Sapphire Rapids” processors. The processors offer the same ISA features as Genoa (-X) and Bergamo, including AVX-512, and support SMT. Overall, the EPYC 8004 series processors are complementary to Genoa (-X) and Bergamo, filling the gap between them and Ryzen for budget-minded and SOHO/SMB server deployments.

Source: Phoronix.

AMD Launches ROCm 5.7

AMD has released ROCm 5.7, the latest version of its GPU compute stack, Phoronix reports. The update includes a new “hipTensor” library, extended support for the ROCgdb debugger with Fortran and OMPD, and optimizations to the rocRAND and MIVisionX libraries. However, there is no official support for new RDNA3 GPUs yet, with speculation that it will be included in ROCm 6.0. The release notes for ROCm 6.0 suggest that it will have fundamental changes and will not be backwards compatible with ROCm 5.x. In terms of supported hardware, only the Radeon VII is officially supported, along with some Radeon Pro and AMD Instinct products. ROCm 5.7 also brings improvements to HIP printf() handling, a beta release of LLVM Address Sanitizer with GPU support, optimizations, and various fixes. More details can be found in the ROCm 5.7 release documentation.

Source: Phoronix.

Impressive Gains for AMD EPYC 9754 “Bergamo” Server Performance in Linux 6.6

Phoronix reports that early testing of the Linux 6.6 kernel shows impressive gains for AMD’s high core count server processors, particularly the EPYC 9754 “Bergamo.” The new Linux 6.6 kernel features the Earliest Eligible Virtual Deadline First (EEVDF) scheduler, which is a major improvement over the existing Completely Fair Scheduler (CFS). The EEVDF scheduler has shown to help with latency sensitive tasks and has led to a lot of CFS code being removed. Additionally, the Linux 6.6 workqueue improvements benefit AMD CPUs and other chiplet-based CPUs with multiple L3 caches. In testing, Linux 6.6 has shown dramatic improvements on higher core count processors like the AMD EPYC 9754.

The article discusses the testing of the Linux 6.6 kernel on an AMD EPYC 9754 server configuration with 128 cores and 256 threads per socket. The benchmarks were carried out using Linux 6.5.1 stable and then with Linux 6.6-rc1. The results showed significant performance improvements with the Linux 6.6 kernel.

Source: Phoronix.

AMD Features Zen 4 and Zen 4c Cores in its New Hybrid Chips

AMD has introduced its first hybrid chips featuring a combination of Zen 4 and smaller Zen 4c CPU cores. This marks the first time AMD has mixed two different types of CPU cores on the same chip. The new architecture is similar to ARM’s big.LITTLE architecture and Intel’s use of Performance and Efficiency cores.

The Zen 4c cores take up less space and have some modifications that may result in lower performance in certain situations. The new chips that feature this hybrid architecture include the Ryzen 3 7440U, Ryzen 5 7540U, and the Ryzen Z1. The Zen 4c cores support hyperthreading, but tend to be paired with fewer graphics compute units and may have less L2 cache memory. A more powerful variant of the Ryzen Z1, the Ryzen Z1 Extreme, is used in the ASUS ROG Ally, which doesn’t use a hybrid chip.

Based on the chip model numbers, it is expected that the new Phoenix 2 processors will deliver lower performance compared to their original Phoenix counterparts. A report analyzing the performance of the Ryzen Z1 processor seems to confirm this.

Source: Liliputing.