Skip to content

fix(inference): swap known-issue precisions back (GB300 FP8, MI355X FP4) #442

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
functionstackx merged 1 commit into from
Jun 11, 2026
Merged

fix(inference): swap known-issue precisions back (GB300 FP8, MI355X FP4) #442

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
22 changes: 11 additions & 11 deletions packages/app/src/lib/known-issues.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,32 +5,32 @@ import { KNOWN_CONFIG_ISSUES, knownIssueCsvNote, matchKnownConfigIssues } from '
const DSR1 = 'DeepSeek-R1-0528';

describe('matchKnownConfigIssues', () => {
it('matches the GB300 Dynamo TRT MTP entry for DeepSeek R1 FP4', () => {
it('matches the GB300 Dynamo TRT MTP entry for DeepSeek R1 FP8', () => {
const issues = matchKnownConfigIssues(DSR1, [
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp4' },
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp8' },
]);
expect(issues).toHaveLength(1);
expect(issues[0].url).toBe('https://github.com/NVIDIA/srt-slurm/issues/51');
});

it('does not match GB300 Dynamo TRT MTP for non-FP4 precisions', () => {
it('does not match GB300 Dynamo TRT MTP for non-FP8 precisions', () => {
const issues = matchKnownConfigIssues(DSR1, [
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp8' },
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp4' },
]);
expect(issues).toHaveLength(0);
});

it('matches the MI355X MoRI SGLang MTP entry for DeepSeek R1 FP8', () => {
it('matches the MI355X MoRI SGLang MTP entry for DeepSeek R1 FP4', () => {
const issues = matchKnownConfigIssues(DSR1, [
{ hwKey: 'mi355x_mori-sglang_mtp', precision: 'fp8' },
{ hwKey: 'mi355x_mori-sglang_mtp', precision: 'fp4' },
]);
expect(issues).toHaveLength(1);
expect(issues[0].url).toBe('https://github.com/sgl-project/sglang/issues/27194');
});

it('does not match MI355X MoRI SGLang MTP for non-FP8 precisions', () => {
it('does not match MI355X MoRI SGLang MTP for non-FP4 precisions', () => {
const issues = matchKnownConfigIssues(DSR1, [
{ hwKey: 'mi355x_mori-sglang_mtp', precision: 'fp4' },
{ hwKey: 'mi355x_mori-sglang_mtp', precision: 'fp8' },
]);
expect(issues).toHaveLength(0);
});
Expand All @@ -54,9 +54,9 @@ describe('matchKnownConfigIssues', () => {

it('returns each issue at most once even with many matching points', () => {
const issues = matchKnownConfigIssues(DSR1, [
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp4' },
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp4' },
{ hwKey: 'mi355x_mori-sglang_mtp', precision: 'fp8' },
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp8' },
{ hwKey: 'gb300_dynamo-trt_mtp', precision: 'fp8' },
{ hwKey: 'mi355x_mori-sglang_mtp', precision: 'fp4' },
]);
expect(issues).toHaveLength(2);
});
Expand Down
4 changes: 2 additions & 2 deletions packages/app/src/lib/known-issues.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -29,7 +29,7 @@ export const KNOWN_CONFIG_ISSUES: KnownConfigIssue[] = [
{
hwKey: 'gb300_dynamo-trt_mtp',
model: Model.DeepSeek_R1,
precisions: ['fp4'],
precisions: ['fp8'],
summary: 'Accuracy issues',
filed: 'Apr 21, 2026',
url: 'https://github.com/NVIDIA/srt-slurm/issues/51',
Expand All @@ -38,7 +38,7 @@ export const KNOWN_CONFIG_ISSUES: KnownConfigIssue[] = [
{
hwKey: 'mi355x_mori-sglang_mtp',
model: Model.DeepSeek_R1,
precisions: ['fp8'],
precisions: ['fp4'],
summary: 'Accuracy issues',
filed: 'Jun 4, 2026',
url: 'https://github.com/sgl-project/sglang/issues/27194',
Expand Down
Loading