Search results

Issue with P2P on PVC

dGPU-Max

Functionality

Status: Open.

#810 In intel/intel-extension-for-pytorch;

· BenBrock opened on Apr 16, 2025

Communication and compute on separate Streams do not overlap

dGPU-Max

Functionality

Status: Open.

#599 In intel/intel-extension-for-pytorch;

· garrett361 opened on Apr 18, 2024

Corruption with Wuerstchen and Stable Cascade models

ARCARC GPUARC GPU

CorrectnessOutput incorrect or unacceptable accuracy lossOutput incorrect or unacceptable accuracy loss

dGPU-Max

Status: Open.

#529 In intel/intel-extension-for-pytorch;

· Disty0 opened on Feb 13, 2024

FP16 output on PVC has randomness

CorrectnessOutput incorrect or unacceptable accuracy lossOutput incorrect or unacceptable accuracy loss

dGPU-Max

LLM

Status: Open.

#526 In intel/intel-extension-for-pytorch;

· rnwang04 opened on Feb 6, 2024

CPU memory leak lora finetuning on XPU

CPUCPU specific issuesCPU specific issues

dGPU-Max

UserExperience

Status: Open.

#462 In intel/intel-extension-for-pytorch;

· yangw1234 opened on Nov 7, 2023

<code>torch.linalg.eigh</code> is significantly slower than expected on Max Series GPU

dGPU-Max

Performance

Status: Open.

#439 In intel/intel-extension-for-pytorch;

· ogrisel opened on Sep 28, 2023

<code>torch.linalg.svd</code> is significantly slower than expected on a Max Series GPU

dGPU-Max

Performance

Status: Open.

#428 In intel/intel-extension-for-pytorch;

· ogrisel opened on Sep 7, 2023

<code>torch.scatter_add_</code> appears to be (about 5x) slower than cuda backend

dGPU-Max

Performance

Status: Open.

#426 In intel/intel-extension-for-pytorch;

· fcharras opened on Sep 6, 2023