Add ECC support for Nvidia professional GPUs - #486
Merged
Merged
Conversation
# Conflicts: # src/interface.c
Only reserve panel width and allocate the ECC window when a monitored GPU actually exposes volatile ECC counters, mirroring the NVLink probe pattern. Consumer GPUs no longer get a widened device header.
The NVLink window was deliberately excluded from device_length(), so it extended past the panel edge. When several devices share a row this overlapped the next device header and was painted over. Factor the line-2 NVLink geometry into shared helpers and include it in the panel width so the allocated window and reserved width cannot drift apart.
Corrected (SBE) errors are routine and can accumulate into the hundreds, while uncorrected (DBE) errors should be zero and signal a fault. Cap corrected at three digits and uncorrected at two, drop the two leading spaces, and size device_ecc for "ECC 999/99" (10 columns).
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds an ECC error counter to the device header, shown right after the power field. NVIDIA professional and datacenter GPUs (e.g. RTX PRO, Quadro, Tesla, A100/H100) expose volatile ECC error counts through NVML; this surfaces them in nvtop.
Display format is
ECC <corrected>/<uncorrected>:The field follows the existing "valid bit" pattern: on GPUs that do not support ECC (consumer cards, older drivers), NVML return
NVML_ERROR_NOT_SUPPORTED, the valid bit stays unset, and the field is simply left blank — no special-casing required.Changes
extract_gpuinfo_common.h: addecc_corrected/ecc_uncorrecteddynamic fields and their valid bitsextract_gpuinfo_nvidia.c: loadnvmlDeviceGetTotalEccErrors(optional symbol, so older NVML libs still work) and query volatile ECC countsinterface.c/interface_internal_common.h: add theecc_infowindow and draw it after the power fieldNotes
ECC 0/0) and an RTX 5090 (no ECC, field correctly hidden).