Bug Report: kvarn6 KV-quantization causes massive performance regression on ROCm (RX 7900 XTX)

Ouverte
#122 14 commentaires 0 réactions 0 personnes assignées Voir sur GitHub

Personne n'a encore pris cette issue.

Évaluation

Difficulté
4/5
Temps estimé
3-5 jours
Accessibilité débutants
30/100
Type d'issue
Bug
Clarté
Plutôt claire
Activité
Active
Stack technique
cpp

Piste de recherche

Reproduce with the two llama-server commands and compare the attached level-5 logs, focusing on where the kvarn6 path diverges from q8_0 in the ROCm backend. Profile prompt processing and token generation separately to find whether the slowdown is in quantization, attention, or kernel dispatch. Done means the regression is localized to a specific ROCm kernel or code path and kvarn6 throughput is no longer an order of magnitude below q8_0.

Rédigé par le modèle d'indexation à partir du texte de l'issue.

Description

Issue: Using KV-quantization kvarn6 results in extremely slow performance compared to q8_0 on AMD ROCm hardware.

Performance Comparison

Both tests used the same prompt (10,410 tokens) and identical settings except for KV quantization.

Note: The kvarn6 run was canceled early due to excessive runtime.

Metric q8_0 kvarn6 Regression
Prompt Processing 821.19 tokens/sec 53.79 tokens/sec ~15.3x slower
Text Generation 29.52 tokens/sec 1.34 tokens/sec ~22x slower

This is a severe performance regression that makes kvarn6 practically unusable.

Steps to Reproduce

  1. Use llama.cpp build 11192 (ee8754c57) with ROCm backend on Windows
  2. Load a large model (tested with Qwen3.6-27B-Fable-Fusion-Q4_K_M)
  3. Run with full offload (-ngl 99) and flash attention (-fa 1)
  4. Compare these two configurations:

q8_0 (baseline):

llama-server.exe -m <model> -ngl 99 -fa 1 -ctk q8_0 -ctv q8_0 -c 40000

kvarn6 (regression):

llama-server.exe -m <model> -ngl 99 -fa 1 -ctk kvarn6 -ctv kvarn6 -c 40000

System Information

  • GPU: AMD Radeon RX 7900 XTX (24 GB VRAM)
  • CPU: AMD Ryzen 9 7900X 12-Core
  • RAM: 64 GB
  • OS: Windows
  • Backend: ROCm (HIP), NO_VMM = 1
  • Build: 11192 (ee8754c57) with Clang 20.1.8 for Windows x86_64
  • Model: Qwen3.6-27B-Fable-Fusion (Q4_K_M, 65 layers, 27.32B params, 17.22 GiB)

Relevant Log Excerpts

q8_0 performance:

prompt processing, n_tokens = 10410, progress = 1.00, t = 11.93 s / 872.56 tokens per second
prompt eval time = 12681.67 ms / 10414 tokens (1.22 ms per token, 821.19 tokens per second)
eval time = 73468.79 ms / 2169 tokens (33.87 ms per token, 29.52 tokens per second)

kvarn6 performance (run canceled after ~5.5 minutes):

prompt processing, n_tokens = 10410, progress = 1.00, t = 193.54 s / 53.79 tokens per second
n_decoded = 144, tg = 1.34 t/s

Expected Behavior

kvarn6 should provide competitive or better performance than q8_0 while using less VRAM, or at least not be an order of magnitude slower.

Actual Behavior

kvarn6 is 15-22x slower than q8_0 on ROCm hardware, making it impractical for use.

Additional Notes

  • This appears to be ROCm-specific; the kvarn6 implementation may not be optimized for AMD GPUs
  • The issue affects both prompt processing and token generation equally severely
  • The kvarn6 run was manually canceled due to excessive runtime; actual degradation may be even more severe over longer generations
  • Full log level 5 logs for both q8_0 and kvarn6 runs are attached to this issue

log_kvarn6_lv5.txt
log_q8_0_lv5.txt

Langage dominant
C++
Étoiles
1.2k
Forks
80
Merge moyen
4 j 20 h
PR mergées (30 j)
2

Préparer son environnement

Ce projet ne fournit ni conteneur de développement, ni Dockerfile, ni guide de contribution : l'installation est à votre charge. Commencez par son README, et consultez notre guide de la première contribution pour les étapes générales.

Par où commencer

  1. Lisez l'issue en entier, puis le guide de contribution du projet.
  2. Signalez en commentaire que vous la prenez — cela évite que deux personnes fassent le même travail.
  3. Forkez le dépôt et travaillez sur une branche.
  4. Ouvrez une pull request qui référence le numéro de l'issue.

Autres issues de Anbeeld/beellama.cpp

Toutes les issues de Anbeeld/beellama.cpp

Issues similaires

Plus d'issues C++

Recevez les nouvelles issues par e-mail

Un résumé court des issues GitHub adaptées aux débutants.