randomllama commited on
Commit
92e6eae
·
verified ·
1 Parent(s): 498eadf

sync: latest measured numbers

Browse files
Files changed (1) hide show
  1. RESULTS.md +10 -0
RESULTS.md CHANGED
@@ -46,3 +46,13 @@ DSA tilelang, NVFP4) DFlash2-on is NOT bit-identical to DFLASH-off at temp 0.**
46
  drafter card's "greedy output matches the target exactly" does not reproduce here; whether
47
  the cause is verify-path numerics on this chip or quant interaction is unresolved. Users
48
  needing bit-exact greedy reproducibility should serve DFLASH-off.
 
 
 
 
 
 
 
 
 
 
 
46
  drafter card's "greedy output matches the target exactly" does not reproduce here; whether
47
  the cause is verify-path numerics on this chip or quant interaction is unresolved. Users
48
  needing bit-exact greedy reproducibility should serve DFLASH-off.
49
+
50
+ ## Thinking mode vs effort (2026-08-28)
51
+ Decode tok/s is IDENTICAL with thinking on or off (~18-22 on this probe; run variance
52
+ exceeds any mode difference). But with a tight max_tokens budget, thinking-ON can spend
53
+ the ENTIRE budget on reasoning and return zero answer: at 600 max_tokens our probe got
54
+ 2,600 chars of reasoning_content and empty content (finish=length), while thinking-OFF
55
+ returned 2,462 chars of pure answer in the same wall time. For agent/tool workloads,
56
+ disable thinking per-request ("chat_template_kwargs": {"enable_thinking": false}) or
57
+ budget max_tokens for reasoning + answer. Conditions: merge-sorted-lists code prompt,
58
+ 600 max_tokens, temp 0, FP8T8V config.