Intel Arc Alchemist challenges the RTX 3070 Ti in the first benchmarks

Intel Arc Alchemist challenges the RTX 3070 Ti in the first benchmarks

The first alleged benchmark results conducted on one of the upcoming Intel Arc Alchemist graphics cards with 512 Vector Engine have been spotted in SiSoftware Sandra's Benchmark Ranker, discovered by well-known leaker @Tum_Apisak. Apparently, the overall GPU Compute score of Intel's upcoming dedicated graphics processing unit is even higher than that of NVIDIA's GeForce RTX 3070 Ti, but a few notes need to be made.

The score, which allegedly refers to the desktop flagship of the Intel Arc Alchemist range, was sent by an unknown person to the database. The GPU has 512 enabled Vector Engines (corresponding to 4,096 Stream Processors), has 4MB of L2 cache and operates at 2.10 GHz. The card comes with 12.8GB of usable memory, which probably indicates there is 16GB of memory in total. Assuming the results presented to SiSoftware Sandra's Benchmark Ranker are true and the drivers are mature enough, the overall score of 9017.52 Mpix / s looks pretty good, as it outperforms NVIDIA's GeForce RTX 3070 Ti, which stops at 8369, 51 Mpix / s.

Arc Alchemist Xe-HPG 512 (MPix / s) GeForce RTX 3070 Ti (MPix / s) FP16 | Half-float GP 35,093 36,510 FP32 | Single-float GP 20,888 27,029 FP64 | Double-float GP 1.000 594 FP128 | Quad-float GP 109 22 Score 9017.52 8369.51 However, when it comes to FP32 compute workloads - precision used most by today's graphics applications - the Intel card performs 23% lower than its competitor based on GA104 GPU. This isn't exactly a spectacular result, at least when it comes to performance in FP32 compute workloads. Obviously, we remind you that the product has not yet been launched, so it could be a prototype not yet optimized at the hardware and / or software level.

Intel Arc Alchemist Xe-512 performs on par with the GeForce RTX 3070 Ti in FP16 compute workloads, which are not widely used for graphics. Additionally, the Xe-HPG GPU outperforms the GA104 by 68% in FP64 compute workloads and nearly 5x (or 395%) in FP128 loads, but neither FP64 nor FP128 are used by graphics applications and are reserved for compute high performance.

if (jQuery ("# ​​crm_srl-th_hardware_d_mh2_1"). is (": visible")) {console.log ("Edinet ADV adding zone: tag crm_srl-th_hardware_d_mh2_1 slot id: th_hardware_d_mh2" ); } Photo Credit: Intel The result published in SiSoftware Sandra's Benchmark Ranker is marked as' Intel (R) Graphics d gfx-driver-ci-master-10002 DCH ReleaseInternal (4096S 512C SM3.0 2.1GHz, 4MB L2, 12.8GB) (OpenCL), 'so we are dealing with a sample and the driver intended for internal testing. In addition, the test was conducted on the 'CoffeeLake Client Platform CannonLake Client System (Intel CoffeeLake S UDIMM RVP)', which is not a gaming system at all.

Powered by Blogger.