Intel’s Ponte Vecchio Xe HPC GPU A0 Silicon Power On, Benchmarked And Detailed

Usman Pirzada

Intel has powered on the Ponte Vecchio A0 silicon and it seems to be working perfectly. This is great news for Intel investors because Ponte Vecchio was the company's most ambitious moonshot in quite a while and considering it had no prior discrete graphics experience it is very impressive that Masooma Bhaiwala and her team, led by Raja Koduri seem to have gotten it right from the get-go.

Intel's Raja Koduri holds the Ponte Vecchio A0 silicon in his hands.

Intel also admitted several faults in its graphics side during Intel Architecture Day 2021 but also said that they had fixed every last one of those with the Xe HPC architecture. The graph below shows a blue line (Intel graphics) and a green line (a competitor that is likely NVIDIA). As you can see, before 2021, the blue line was significantly under the green line but as of the Ponte Vecchio A0 silicon power on, this has changed.

Related StoryUsman Pirzada
Intel Max Series Data Center GPU: 128GB HBM2e, 52 TFLOPs Per OAM, 8 OAM Max, Up To 2x Faster Than NVIDIA A100 In Specific Workloads

Intel revealed that building Ponte Vecchio had required a recreation and retooling of basically *everything* involved in the design process.

The Ponte Vecchio package consists of 47 tiles that are built on different foundries (some in Intel, some in TSMC) and require a ton of packaging technologies to come together as a cohesive whole.

The team at Folsom is measuring 45 TFLOPs of FP32 throughput, 5 TBps of memory fabric bandwidth and 2 TBps of connectivity bandwidth validating that all parts of the A0 silicon are working correctly and track to be a major breakthrough in graphics.

No demo is ever complete without benchmarks and Intel even threw in some Resnet benchmarks that show that Intel has broken through the world record score with the A0 Ponte Vecchio silicon:

Intel's Xe HPC will exist in 1 stack and 2 stack designs and Xe Link will be used to couple Ponte Vecchio's in multiple subsystems.

intel-architecture-day-2021_pressdeck_final_embargo-compressed-163
intel-architecture-day-2021_pressdeck_final_embargo-compressed-164
intel-architecture-day-2021_pressdeck_final_embargo-compressed-180

Unlike Xe HPG, an Xe HPC core in the Ponte Vecchio GPU has 8 vector engines (512 bit) and 8 matrix engines (4096 bit).

It is clear that Intel appears to have overcome the valley of stagnation and is making significant progress in leaps and bounds when it comes to high performance graphics. The Aurora super computer will utilize Ponte Vecchio GPUs and the company showed off a single Aurora Blade which will constitute one computational unit of the supercomputing cluster.

The rest of the Xe HPC and Ponte Vecchio presentation is given below:

intel-architecture-day-2021_pressdeck_final_embargo-compressed-152
intel-architecture-day-2021_pressdeck_final_embargo-compressed-190
intel-architecture-day-2021_pressdeck_final_embargo-compressed-189
intel-architecture-day-2021_pressdeck_final_embargo-compressed-188
intel-architecture-day-2021_pressdeck_final_embargo-compressed-187
intel-architecture-day-2021_pressdeck_final_embargo-compressed-186
intel-architecture-day-2021_pressdeck_final_embargo-compressed-185
intel-architecture-day-2021_pressdeck_final_embargo-compressed-184
intel-architecture-day-2021_pressdeck_final_embargo-compressed-183
intel-architecture-day-2021_pressdeck_final_embargo-compressed-182
intel-architecture-day-2021_pressdeck_final_embargo-compressed-181
intel-architecture-day-2021_pressdeck_final_embargo-compressed-180
intel-architecture-day-2021_pressdeck_final_embargo-compressed-179
intel-architecture-day-2021_pressdeck_final_embargo-compressed-178
intel-architecture-day-2021_pressdeck_final_embargo-compressed-177
intel-architecture-day-2021_pressdeck_final_embargo-compressed-176
intel-architecture-day-2021_pressdeck_final_embargo-compressed-175
intel-architecture-day-2021_pressdeck_final_embargo-compressed-174
intel-architecture-day-2021_pressdeck_final_embargo-compressed-173
intel-architecture-day-2021_pressdeck_final_embargo-compressed-172
intel-architecture-day-2021_pressdeck_final_embargo-compressed-171
intel-architecture-day-2021_pressdeck_final_embargo-compressed-170
intel-architecture-day-2021_pressdeck_final_embargo-compressed-169
intel-architecture-day-2021_pressdeck_final_embargo-compressed-168
intel-architecture-day-2021_pressdeck_final_embargo-compressed-167
intel-architecture-day-2021_pressdeck_final_embargo-compressed-166
intel-architecture-day-2021_pressdeck_final_embargo-compressed-165
intel-architecture-day-2021_pressdeck_final_embargo-compressed-164
intel-architecture-day-2021_pressdeck_final_embargo-compressed-163
intel-architecture-day-2021_pressdeck_final_embargo-compressed-162
intel-architecture-day-2021_pressdeck_final_embargo-compressed-161
intel-architecture-day-2021_pressdeck_final_embargo-compressed-160
intel-architecture-day-2021_pressdeck_final_embargo-compressed-159
intel-architecture-day-2021_pressdeck_final_embargo-compressed-158
intel-architecture-day-2021_pressdeck_final_embargo-compressed-157
intel-architecture-day-2021_pressdeck_final_embargo-compressed-156
intel-architecture-day-2021_pressdeck_final_embargo-compressed-155
intel-architecture-day-2021_pressdeck_final_embargo-compressed-154
intel-architecture-day-2021_pressdeck_final_embargo-compressed-153

Deal of the Day

Comments