48-Core Sapphire Rapids Xeons Match AMD’s Zen 3 Flagships in New Benchmarks
A pair of prototype Intel Xeon Platinum 8468 Sapphire Rapids 48-core server processors were recently tested in the United States. geek bench 5, As shared by @BenchLeaks on TwitterThey feature some impressive performance numbers that match AMD’s best twin EPYC 64-core config in the same benchmarks, and these chips aren’t even the 56-core variant of the flagship.
The shared Sapphire Rapids system featured twin Xeon Platinum 8468 CPUs with a total of 96 cores and 192 threads. Combining the two chips yielded a score of 1,257 points single-threaded and 74,586 points multi-threaded.
One of the highest performing dual AMD EPYC processor configurations we can find on Geekbench 5 for comparison features Twin AMD EPYC 7763 processors The total core count is 128 cores and 256 threads. The two chips scored 1249 points in the single-threaded test and 75,539 points in the multi-threaded test.
This gives a difference of 1% for both single-core and multi-core on both platforms, effectively tying them together in comparable performance. But when looking at the individual performance of each core during multithreaded workloads, Intel is 23% faster per core compared to AMD.
This is important because single-threaded benchmarks show maximum performance for an individual core only when all other cores are idle, allowing that single core to use its extra power and heat reserves. It is important to As a result, individual core performance under load shows what per-core performance looks like under maximum load.
Note, however, that Geekbench 5 is not particularly useful as a benchmark for real-world performance results. So take this data with a grain of salt. On the contrary, Geekbench at least gives us a general idea of Sapphire Rapids performance.
A quick recap on Sapphire Rapids and its delays
Sapphire Rapids is Intel’s upcoming server architecture which is designed to replace Intel’s current 14nm Cascade Lake Xeon lineup. As a result, Sapphire Rapids will feature some serious upgrades including Intel’s leading Golden Cove cores running on the Intel 7 (10nm Enhanced SuperFin) node – similar to Intel’s current Alder Lake parts.
Like Alder Lake, it supports PCIe 5.0 and DDR5 (some parts are HBM compatible). However, unlike Alder Lake, efficiency cores are not supported. From what we know, Sapphire Rapids will feature up to 56 cores. But thanks to multi-socket motherboards, the core count can grow to 112 or more. Sapphire Rapids also has additional features like his AMX and Intel DSA specifically for server architectures.
Unfortunately, Sapphire Rapids has allegedly turned into a bug-ridden nightmare for Intel, with multiple delays pushing back the platform’s release window. Originally, the Sapphire Rapids was meant to face off against AMD’s Zen 3 EPYC Milan processors (such as the one featured in this article), but thanks to 500 or so bugs found in the microarchitecture, the launch window has been dwarfed by his Moved to early 2023.
This means Sapphire Rapids will have to compete with current Zen 3 parts and AMD’s future Zen 4 EPYC processors (codenamed Genoa) in 2023. Thankfully, it looks like Intel’s future server architecture can at least compete with and slightly outperform AMD’s Zen 3 competitor. But we don’t know how it will perform against AMD’s future Zen 4 architecture, which has already been released in the form of Ryzen 7000.