请教各位大神，5090 上 stara 如何支持多线程能快点啊？

[tester12340]

另外，还可以把他们拉到一起 放到 telegrame 里开会 * T_T*。

重大发现 @perter 你真的是太牛逼了

loading the model (the first start takes a minute or two) ...
[strata] starting the engine: reading the model's weights ...
[strata] loading the experts the GPU does not hold into RAM (of about 38 GB) and locking part of them for the GPU.
YOUR PC CAN BE SLOW OR STOP RESPONDING FOR 1-3 MINUTES NOW - this is normal.
Please wait and don't close this window; the browser opens when it is ready.
[strata] filling the GPU's expert cache (11869 experts, 15.28 GiB of VRAM) ...
[strata] almost ready ...
ready: http://192.168.0.3:8080/v1 (OpenAI: /v1/chat/completions, Anthropic: /v1/messages, context 131072 tokens)
open http://192.168.0.3:8080/ in a browser to chat; close this window to stop the model
from other devices: http://192.168.0.3:8080/ (API: http://192.168.0.3:8080/v1)
from other devices: http://172.18.224.1:8080/ (API: http://172.18.224.1:8080/v1)
from other devices: http://172.28.224.1:8080/ (API: http://172.28.224.1:8080/v1)
WARNING: no API key - anyone on your network can use this model. Add "api_key": "..." to the config (clients send it as their API key; the web page asks for it)
nothing arrives? Windows Firewall blocks it until allowed: accept its prompt for Python, or run in an admin PowerShell:
New-NetFirewallRule -DisplayName "Strata 8080" -Direction Inbound -Protocol TCP -LocalPort 8080 -Action Allow -Profile Private
(and set this network to Private in Windows' network settings)
[strata] reading the prompt: 8,192 of 63,400 tokens, 3 s so far
[strata] reading the prompt: 13,362 of 63,400 tokens, 6 s so far
[strata] reading the prompt: 21,554 of 63,400 tokens, 10 s so far
[strata] reading the prompt: 143 of 148 tokens, 6 s so far
[strata] reading the prompt: 29,746 of 63,400 tokens, 15 s so far
[strata] thinking: 3 of max 130916 tokens, 0.9 tok/s, 10 s
[strata] done: 43 tokens in 10 s (10.9 tok/s) (stop, cancel=False), expert cache 76.0% hit
[strata] reading the prompt: 37,938 of 63,400 tokens, 18 s so far
[strata] reading the prompt: 46,130 of 63,400 tokens, 21 s so far
[strata] reading the prompt: 54,322 of 63,400 tokens, 24 s so far
[strata] reading the prompt: 62,514 of 63,400 tokens, 27 s so far
[strata] reading the prompt: 63,395 of 63,400 tokens, 30 s so far
[strata] thinking: 2 of max 65536 tokens, 3.5 tok/s, 31 s
[strata] reading the prompt: 162 of 167 tokens, 4 s so far
[strata] thinking: 3 of max 65536 tokens, 1.9 tok/s, 32 s
[strata] done: 40 tokens in 5 s (58.2 tok/s) (stop, cancel=False), expert cache 75.8% hit
[strata] thinking: 45 of max 65536 tokens, 16.4 tok/s, 33 s
[strata] answering: 137 of max 65536 tokens, 36.3 tok/s, 34 s
[strata] done: 172 tokens in 35 s (41.8 tok/s) (stop, cancel=False), expert cache 77.2% hit
[strata] response_format json_schema: the Python package jsonschema is not installed, so answers are only checked to be one JSON object (python -m pip install "jsonschema>=4.23,<5")
[strata] reading the prompt: 332 of 339 tokens, 2 s so far
[strata] reading the prompt: 8,192 of 16,007 tokens, 5 s so far
[strata] answering: 4 of max 130725 tokens, 1.3 tok/s, 5 s
[strata] done: 8 tokens in 5 s (2.5 tok/s) (stop, cancel=False), expert cache 63.7% hit
[strata] reading the prompt: 15,994 of 16,007 tokens, 8 s so far
[strata] answering: 5 of max 65536 tokens, 12.7 tok/s, 9 s
[strata] answering: 117 of max 65536 tokens, 83.7 tok/s, 10 s
[strata] done: 156 tokens in 11 s (88.4 tok/s) (stop, cancel=False), expert cache 90.1% hit
[strata] answering: 72 of max 130878 tokens, 99.0 tok/s, 1 s
[strata] done: 84 tokens in 1 s (100.4 tok/s) (stop, cancel=False), expert cache 94.4% hit
[strata] reading the prompt: 63,532 of 63,537 tokens, 1 s so far
[strata] done: 110 tokens in 2 s (129.2 tok/s) (stop, cancel=False), expert cache 96.7% hit
[strata] reading the prompt: 16,169 of 16,176 tokens, 1 s so far
[strata] done: 75 tokens in 2 s (128.9 tok/s) (stop, cancel=False), expert cache 94.1% hit
[strata] thinking: 53 of max 130826 tokens, 109.6 tok/s, 1 s
[strata] done: 76 tokens in 1 s (116.6 tok/s) (stop, cancel=False), expert cache 95.4% hit
[strata] writing a tool call: terminal: 62 of max 65536 tokens, 99.9 tok/s, 1 s
[strata] writing a tool call: terminal: 171 of max 65536 tokens, 104.2 tok/s, 2 s
[strata] done: 214 tokens in 2 s (107.6 tok/s) (stop, cancel=False), expert cache 81.0% hit
[strata] reading the prompt: 685 of 692 tokens, 1 s so far
[strata] done: 4 tokens in 2 s (235.1 tok/s) (stop, cancel=False), expert cache 89.9% hit
[strata] reading the prompt: 18,004 of 18,011 tokens, 3 s so far
[strata] writing a tool call: search_files: 104 of max 65536 tokens, 131.0 tok/s, 4 s
[strata] done: 165 tokens in 4 s (133.9 tok/s) (stop, cancel=False), expert cache 94.0% hit
[strata] reading the prompt: 18,581 of 18,588 tokens, 1 s so far
[strata] done: 101 tokens in 2 s (144.6 tok/s) (stop, cancel=False), expert cache 92.3% hit
[strata] writing a tool call: terminal: 134 of max 65536 tokens, 154.8 tok/s, 1 s
[strata] done: 152 tokens in 1 s (153.7 tok/s) (stop, cancel=False), expert cache 94.2% hit
[strata] reading the prompt: 19,179 of 19,186 tokens, 1 s so far
[strata] done: 140 tokens in 2 s (170.5 tok/s) (stop, cancel=False), expert cache 94.4% hit
[strata] reading the prompt: 19,731 of 19,738 tokens, 1 s so far
[strata] done: 30 tokens in 2 s (181.3 tok/s) (stop, cancel=False), expert cache 94.2% hit
[strata] reading the prompt: 20,373 of 20,380 tokens, 2 s so far
[strata] done: 114 tokens in 2 s (147.1 tok/s) (stop, cancel=False), expert cache 92.2% hit
[strata] writing a tool call: execute_code: 36 of max 65536 tokens, 162.0 tok/s, 1 s
[strata] done: 158 tokens in 2 s (173.4 tok/s) (stop, cancel=False), expert cache 95.7% hit
[strata] writing a tool call: execute_code: 40 of max 65536 tokens, 197.4 tok/s, 1 s
[strata] done: 146 tokens in 2 s (170.8 tok/s) (stop, cancel=False), expert cache 95.6% hit
[strata] reading the prompt: 21,781 of 21,788 tokens, 2 s so far
[strata] writing a tool call: execute_code: 130 of max 65536 tokens, 151.6 tok/s, 3 s
[strata] done: 183 tokens in 3 s (152.0 tok/s) (stop, cancel=False), expert cache 95.8% hit
[strata] reading the prompt: 23,252 of 23,259 tokens, 3 s so far
[strata] reading the prompt: 257 of 262 tokens, 3 s so far
[strata] reading the prompt: 4 of max 65536 tokens, 3.5 tok/s, 4 s
[strata] done: 27 tokens in 4 s (54.7 tok/s) (stop, cancel=False), expert cache 93.3% hit
[strata] writing a tool call: execute_code: 92 of max 65536 tokens, 42.7 tok/s, 5 s
[strata] done: 198 tokens in 6 s (71.2 tok/s) (stop, cancel=False), expert cache 96.7% hit
[strata] writing a tool call: execute_code: 18 of max 65536 tokens, 209.1 tok/s, 1 s
[strata] done: 147 tokens in 2 s (188.9 tok/s) (stop, cancel=False), expert cache 96.0% hit
[strata] reading the prompt: 393 of 400 tokens, 1 s so far
[strata] done: 4 tokens in 1 s (173.7 tok/s) (stop, cancel=False), expert cache 84.4% hit
[strata] reading the prompt: 24,031 of 24,038 tokens, 2 s so far
[strata] writing a tool call: execute_code: 149 of max 65536 tokens, 170.6 tok/s, 3 s
[strata] done: 178 tokens in 3 s (176.1 tok/s) (stop, cancel=False), expert cache 96.3% hit
[strata] writing a tool call: execute_code: 10 of max 65536 tokens, 167.9 tok/s, 1 s
[strata] writing a tool call: execute_code: 203 of max 65536 tokens, 188.2 tok/s, 2 s
[strata] done: 319 tokens in 3 s (186.0 tok/s) (stop, cancel=False), expert cache 96.2% hit
[strata] reading the prompt: 25,276 of 25,283 tokens, 2 s so far
[strata] writing a tool call: execute_code: 150 of max 65536 tokens, 171.4 tok/s, 3 s
[strata] done: 281 tokens in 3 s (181.6 tok/s) (stop, cancel=False), expert cache 95.0% hit
[strata] reading the prompt: 26,374 of 26,381 tokens, 1 s so far
[strata] answering: 70 of max 65536 tokens, 82.1 tok/s, 3 s
[strata] answering: 185 of max 65536 tokens, 99.6 tok/s, 4 s
[strata] answering: 286 of max 65536 tokens, 99.6 tok/s, 5 s
[strata] done: 331 tokens in 5 s (102.1 tok/s) (stop, cancel=False), expert cache 88.3% hit
[strata] reading the prompt: 28,957 of 28,964 tokens, 3 s so far
[strata] done: 43 tokens in 3 s (88.8 tok/s) (stop, cancel=False), expert cache 80.0% hit
[strata] reading the prompt: 32,102 of 32,109 tokens, 3 s so far
[strata] done: 67 tokens in 4 s (108.6 tok/s) (stop, cancel=False), expert cache 84.8% hit
[strata] reading the prompt: 32,783 of 32,790 tokens, 2 s so far
[strata] writing a tool call: read_file: 108 of max 65536 tokens, 120.1 tok/s, 3 s
[strata] done: 111 tokens in 3 s (123.4 tok/s) (stop, cancel=False), expert cache 84.6% hit
[strata] reading the prompt: 33,542 of 33,549 tokens, 2 s so far
[strata] writing a tool call: execute_code: 116 of max 65536 tokens, 126.6 tok/s, 3 s
[strata] done: 173 tokens in 4 s (128.8 tok/s) (stop, cancel=False), expert cache 89.3% hit
[strata] reading the prompt: 34,115 of 34,122 tokens, 1 s so far
[strata] writing a tool call: execute_code: 120 of max 65536 tokens, 136.6 tok/s, 2 s
[strata] done: 176 tokens in 3 s (142.8 tok/s) (stop, cancel=False), expert cache 93.2% hit
[strata] reading the prompt: 34,861 of 34,868 tokens, 2 s so far
[strata] writing a tool call: execute_code: 115 of max 65536 tokens, 131.4 tok/s, 3 s
[strata] done: 192 tokens in 3 s (135.5 tok/s) (stop, cancel=False), expert cache 93.2% hit
[strata] reading the prompt: 35,986 of 35,993 tokens, 2 s so far
[strata] answering: 81 of max 65536 tokens, 94.2 tok/s, 3 s
[strata] done: 175 tokens in 4 s (115.1 tok/s) (stop, cancel=False), expert cache 91.1% hit
[strata] reading the prompt: 37,260 of 37,267 tokens, 3 s so far
[strata] answering: 118 of max 65536 tokens, 143.9 tok/s, 4 s
[strata] writing a tool call: skill_manage: 260 of max 65536 tokens, 141.7 tok/s, 5 s
[strata] writing a tool call: skill_manage: 380 of max 65536 tokens, 132.0 tok/s, 6 s
[strata] writing a tool call: skill_manage: 486 of max 65536 tokens, 124.9 tok/s, 7 s
[strata] writing a tool call: skill_manage: 628 of max 65536 tokens, 128.1 tok/s, 8 s
[strata] done: 635 tokens in 8 s (128.3 tok/s) (stop, cancel=False), expert cache 89.8% hit
[strata] writing a tool call: skill_manage: 123 of max 65536 tokens, 157.8 tok/s, 1 s
[strata] writing a tool call: skill_manage: 311 of max 65536 tokens, 173.4 tok/s, 2 s
[strata] writing a tool call: skill_manage: 469 of max 65536 tokens, 167.8 tok/s, 3 s
[strata] done: 498 tokens in 3 s (168.8 tok/s) (stop, cancel=False), expert cache 87.2% hit
[strata] reading the prompt: 38,509 of 38,516 tokens, 2 s so far
[strata] answering: 68 of max 65536 tokens, 78.5 tok/s, 3 s
[strata] writing a tool call: skill_manage: 200 of max 65536 tokens, 106.5 tok/s, 4 s
[strata] writing a tool call: skill_manage: 305 of max 65536 tokens, 105.5 tok/s, 5 s
[strata] done: 358 tokens in 5 s (112.3 tok/s) (stop, cancel=False), expert cache 86.8% hit
[strata] answering: 71 of max 65536 tokens, 101.8 tok/s, 1 s
[strata] answering: 192 of max 65536 tokens, 112.3 tok/s, 2 s
[strata] done: 199 tokens in 2 s (110.9 tok/s) (stop, cancel=False), expert cache 93.1% hit
[strata] reading the prompt: 29,114 of 29,121 tokens, 3 s so far
[strata] writing a tool call: terminal: 53 of max 65536 tokens, 84.2 tok/s, 4 s
[strata] writing a tool call: execute_code: 177 of max 65536 tokens, 108.4 tok/s, 5 s
[strata] writing a tool call: execute_code: 326 of max 65536 tokens, 122.9 tok/s, 6 s
[strata] writing a tool call: execute_code: 454 of max 65536 tokens, 122.8 tok/s, 7 s
[strata] done: 535 tokens in 8 s (126.8 tok/s) (stop, cancel=False), expert cache 86.1% hit
[strata] reading the prompt: 30,449 of 30,456 tokens, 2 s so far
[strata] writing a tool call: execute_code: 134 of max 65536 tokens, 163.2 tok/s, 3 s
[strata] writing a tool call: execute_code: 320 of max 65536 tokens, 174.9 tok/s, 4 s
[strata] done: 376 tokens in 4 s (171.9 tok/s) (stop, cancel=False), expert cache 94.8% hit
[strata] reading the prompt: 32,621 of 32,628 tokens, 3 s so far
[strata] writing a tool call: execute_code: 100 of max 65536 tokens, 127.9 tok/s, 4 s
[strata] writing a tool call: execute_code: 273 of max 65536 tokens, 152.8 tok/s, 5 s
[strata] writing a tool call: execute_code: 431 of max 65536 tokens, 154.5 tok/s, 6 s
[strata] done: 459 tokens in 6 s (157.4 tok/s) (stop, cancel=False), expert cache 94.7% hit
[strata] reading the prompt: 34,604 of 34,611 tokens, 3 s so far
[strata] writing a tool call: execute_code: 137 of max 65536 tokens, 172.6 tok/s, 4 s
[strata] writing a tool call: execute_code: 286 of max 65536 tokens, 158.4 tok/s, 5 s
[strata] done: 380 tokens in 5 s (164.6 tok/s) (stop, cancel=False), expert cache 95.0% hit
[strata] reading the prompt: 2 of max 65536 tokens, 37.4 tok/s, 1 s
[strata] writing a tool call: execute_code: 182 of max 65536 tokens, 170.3 tok/s, 2 s
[strata] done: 254 tokens in 2 s (170.7 tok/s) (stop, cancel=False), expert cache 95.0% hit
[strata] reading the prompt: 434 of 441 tokens, 1 s so far
[strata] done: 4 tokens in 1 s (67.7 tok/s) (stop, cancel=False), expert cache 77.5% hit
[strata] reading the prompt: 31,482 of 35,584 tokens, 3 s so far
[strata] reading the prompt: 35,577 of 35,584 tokens, 6 s so far
[strata] writing a tool call: execute_code: 88 of max 65536 tokens, 193.0 tok/s, 7 s
[strata] writing a tool call: execute_code: 250 of max 65536 tokens, 168.9 tok/s, 8 s
[strata] done: 268 tokens in 8 s (168.1 tok/s) (stop, cancel=False), expert cache 94.0% hit
[strata] reading the prompt: 452 of 459 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (181.7 tok/s) (stop, cancel=False), expert cache 83.0% hit
[strata] reading the prompt: 36,015 of 36,016 tokens, 1 s so far
[strata] writing a tool call: execute_code: 189 of max 65536 tokens, 188.2 tok/s, 2 s
[strata] done: 274 tokens in 3 s (185.0 tok/s) (stop, cancel=False), expert cache 93.9% hit
[strata] reading the prompt: 458 of 465 tokens, 1 s so far
[strata] done: 4 tokens in 2 s (71.4 tok/s) (stop, cancel=False), expert cache 84.2% hit
[strata] reading the prompt: 36,454 of 36,455 tokens, 1 s so far
[strata] writing a tool call: execute_code: 192 of max 65536 tokens, 193.2 tok/s, 2 s
[strata] done: 269 tokens in 3 s (184.7 tok/s) (stop, cancel=False), expert cache 93.1% hit
[strata] reading the prompt: 453 of 460 tokens, 1 s so far
[strata] done: 4 tokens in 2 s (70.7 tok/s) (stop, cancel=False), expert cache 83.6% hit
[strata] reading the prompt: 36,881 of 36,888 tokens, 1 s so far
[strata] writing a tool call: execute_code: 168 of max 65536 tokens, 185.1 tok/s, 2 s
[strata] done: 269 tokens in 3 s (181.6 tok/s) (stop, cancel=False), expert cache 92.7% hit
[strata] reading the prompt: 453 of 460 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (69.5 tok/s) (stop, cancel=False), expert cache 82.8% hit
[strata] reading the prompt: 37,299 of 37,300 tokens, 1 s so far
[strata] writing a tool call: execute_code: 184 of max 65536 tokens, 185.8 tok/s, 2 s
[strata] done: 269 tokens in 3 s (174.9 tok/s) (stop, cancel=False), expert cache 92.4% hit
[strata] reading the prompt: 453 of 460 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (163.1 tok/s) (stop, cancel=False), expert cache 82.0% hit
[strata] reading the prompt: 37,727 of 37,734 tokens, 1 s so far
[strata] writing a tool call: execute_code: 165 of max 65536 tokens, 177.9 tok/s, 2 s
[strata] done: 269 tokens in 3 s (173.6 tok/s) (stop, cancel=False), expert cache 91.9% hit
[strata] reading the prompt: 453 of 460 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (181.6 tok/s) (stop, cancel=False), expert cache 81.2% hit
[strata] reading the prompt: 38,158 of 38,165 tokens, 1 s so far
[strata] writing a tool call: execute_code: 166 of max 65536 tokens, 185.7 tok/s, 2 s
[strata] done: 272 tokens in 3 s (177.2 tok/s) (stop, cancel=False), expert cache 92.3% hit
[strata] reading the prompt: 456 of 463 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (181.7 tok/s) (stop, cancel=False), expert cache 81.2% hit
[strata] reading the prompt: 38,573 of 38,580 tokens, 1 s so far
[strata] writing a tool call: execute_code: 161 of max 65536 tokens, 181.1 tok/s, 2 s
[strata] done: 271 tokens in 3 s (175.3 tok/s) (stop, cancel=False), expert cache 91.5% hit
[strata] reading the prompt: 455 of 462 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (159.8 tok/s) (stop, cancel=False), expert cache 80.6% hit
[strata] reading the prompt: 38,987 of 38,994 tokens, 1 s so far
[strata] writing a tool call: execute_code: 165 of max 65536 tokens, 178.7 tok/s, 2 s
[strata] done: 271 tokens in 3 s (170.4 tok/s) (stop, cancel=False), expert cache 91.3% hit
[strata] reading the prompt: 455 of 462 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (63.4 tok/s) (stop, cancel=False), expert cache 81.1% hit
[strata] reading the prompt: 39,400 of 39,407 tokens, 1 s so far
[strata] writing a tool call: execute_code: 165 of max 65536 tokens, 183.1 tok/s, 2 s
[strata] done: 272 tokens in 3 s (167.5 tok/s) (stop, cancel=False), expert cache 91.0% hit
[strata] reading the prompt: 456 of 463 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (170.0 tok/s) (stop, cancel=False), expert cache 79.7% hit
[strata] reading the prompt: 39,815 of 39,822 tokens, 1 s so far
[strata] writing a tool call: execute_code: 165 of max 65536 tokens, 175.7 tok/s, 2 s
[strata] done: 271 tokens in 3 s (170.7 tok/s) (stop, cancel=False), expert cache 90.9% hit
[strata] reading the prompt: 455 of 462 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (58.3 tok/s) (stop, cancel=False), expert cache 80.2% hit
[strata] reading the prompt: 40,229 of 40,236 tokens, 1 s so far
[strata] writing a tool call: execute_code: 169 of max 65536 tokens, 181.7 tok/s, 2 s
[strata] done: 271 tokens in 3 s (176.4 tok/s) (stop, cancel=False), expert cache 90.8% hit
[strata] reading the prompt: 455 of 462 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (71.4 tok/s) (stop, cancel=False), expert cache 80.0% hit
[strata] reading the prompt: 40,643 of 40,650 tokens, 1 s so far
[strata] writing a tool call: execute_code: 162 of max 65536 tokens, 184.2 tok/s, 2 s
[strata] done: 272 tokens in 3 s (177.6 tok/s) (stop, cancel=False), expert cache 91.0% hit
[strata] reading the prompt: 456 of 463 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (66.1 tok/s) (stop, cancel=False), expert cache 79.4% hit
[strata] reading the prompt: 41,058 of 41,065 tokens, 1 s so far
[strata] writing a tool call: execute_code: 157 of max 65536 tokens, 170.7 tok/s, 2 s
[strata] done: 273 tokens in 3 s (162.0 tok/s) (stop, cancel=False), expert cache 90.8% hit
[strata] reading the prompt: 457 of 464 tokens, 2 s so far
[strata] done: 4 tokens in 2 s (66.1 tok/s) (stop, cancel=False), expert cache 78.3% hit
[strata] reading the prompt: 41,474 of 41,481 tokens, 1 s so far
[strata] writing a tool call: execute_code: 159 of max 65536 tokens, 177.6 tok/s, 2 s
[strata] done: 277 tokens in 3 s (164.0 tok/s) (stop, cancel=False), expert cache 90.2% hit

现在多开，来回切agents 爽到极致啊， 我感觉比 27b 都快，而且现在这工作质量真的太棒了！ 感谢锤子兄弟提供这么好的地方，感谢peter 感谢论坛里的大神们。

不过现在问题来了，我刚订的 max x5 ultra 96 是不是应该退掉，是不是会被完败，请大神们赐教 Thanks♪(･ω･)ﾉ

[perter]

@tester12340
发一个我48GB 的 rtx pro 5000的全套参数给你参考，你根据自己内存大小调整。

"args": [
  "--pack", "C:\\work\\Strata-data\\packs\\iq3_xxs",
  "--native", "C:\\work\\Strata-data\\models\\IQ3_XXS\\Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00001-of-00002.gguf",
  "--ple-gguf", "C:\\work\\Strata-data\\models\\IQ3_XXS\\Qwen3.8-Flash-Next-GSQ-RCO-IQ3_XXS-00002-of-00002.gguf",
  "--expert-profile", "C:\\work\\Strata\\data\\expert-profile.bin",
  "--resident-experts",
  "--expert-cache", "19000",
  "--prefill", "auto",
  "--spec", "4",
  "--mtp", "C:\\work\\Strata-data\\mtp\\rt",
  "--max-context", "262144",
  "--kv", "k8v4",
  "--vision", 
  "--vram-reserve-mib", "3400", 
  "--pcie-frac", "0.00", 
  "--spec-min-p", "0.70",
  "--conversation-cache-mib", "27648",
  "--conversation-cache-slots", "9",
  "--conversation-cache-min-free-mib", "8192",
  "--prompt-cache-every", "8192"
 ],

[tester12340]

跟我逾期的一样。就是短上下文没问题。但长上下文就会进入无尽的循环。 然后还会自动降级到 65535 。 32gb已经被榨干。我想是 32gb 就是老黄这个奸商设计的短板吧。 还是 48g香啊

trata] reading the prompt: 1,024 of 1,031 tokens, 3 s so far
[strata] done: 4 tokens in 3 s (56.3 tok/s) (stop, cancel=False), expert cache 56.0% hit
[strata] reading the prompt: 8,192 of 8,975 tokens, 3 s so far
[strata] reading the prompt: 8,968 of 8,975 tokens, 5 s so far
[strata] answering: 40 of max 122089 tokens, 95.7 tok/s, 6 s
[strata] answering: 163 of max 122089 tokens, 111.8 tok/s, 7 s
[strata] answering: 287 of max 122089 tokens, 114.3 tok/s, 8 s
[strata] answering: 418 of max 122089 tokens, 117.5 tok/s, 9 s
[strata] answering: 572 of max 122089 tokens, 125.2 tok/s, 10 s
[strata] answering: 717 of max 122089 tokens, 128.6 tok/s, 11 s
[strata] answering: 842 of max 122089 tokens, 127.8 tok/s, 12 s
[strata] answering: 1000 of max 122089 tokens, 131.5 tok/s, 13 s
[strata] answering: 1161 of max 122089 tokens, 134.9 tok/s, 14 s
[strata] answering: 1344 of max 122089 tokens, 139.9 tok/s, 15 s
[strata] answering: 1497 of max 122089 tokens, 140.9 tok/s, 16 s
[strata] answering: 1658 of max 122089 tokens, 142.5 tok/s, 17 s
[strata] answering: 1817 of max 122089 tokens, 143.7 tok/s, 18 s
[strata] answering: 1981 of max 122089 tokens, 145.1 tok/s, 19 s
[strata] answering: 2164 of max 122089 tokens, 147.5 tok/s, 20 s
[strata] answering: 2309 of max 122089 tokens, 146.7 tok/s, 21 s
[strata] thinking: 19 of max 130783 tokens, 41.2 tok/s, 1 s
[strata] answering: 2337 of max 122089 tokens, 139.4 tok/s, 23 s
[strata] answering: 77 of max 130783 tokens, 52.2 tok/s, 2 s
[strata] done: 82 tokens in 2 s (52.4 tok/s) (stop, cancel=False), expert cache 59.8% hit
[strata] answering: 2394 of max 122089 tokens, 133.5 tok/s, 24 s
[strata] answering: 2553 of max 122089 tokens, 134.8 tok/s, 25 s
[strata] answering: 2713 of max 122089 tokens, 135.9 tok/s, 26 s
[strata] answering: 2879 of max 122089 tokens, 137.2 tok/s, 27 s
[strata] answering: 3050 of max 122089 tokens, 138.6 tok/s, 28 s
[strata] answering: 3236 of max 122089 tokens, 140.6 tok/s, 29 s
[strata] answering: 3414 of max 122089 tokens, 142.0 tok/s, 30 s
[strata] answering: 3588 of max 122089 tokens, 143.2 tok/s, 31 s
[strata] answering: 3783 of max 122089 tokens, 145.1 tok/s, 32 s
[strata] answering: 3975 of max 122089 tokens, 146.8 tok/s, 33 s
[strata] answering: 4161 of max 122089 tokens, 148.2 tok/s, 34 s
[strata] answering: 4344 of max 122089 tokens, 149.3 tok/s, 35 s
[strata] answering: 4543 of max 122089 tokens, 150.9 tok/s, 36 s
[strata] answering: 4755 of max 122089 tokens, 152.7 tok/s, 37 s
[strata] answering: 4927 of max 122089 tokens, 153.2 tok/s, 38 s
[strata] answering: 5113 of max 122089 tokens, 154.0 tok/s, 39 s
[strata] done: 5184 tokens in 39 s (154.4 tok/s) (stop, cancel=False), expert cache 90.8% hit
[strata] reading the prompt: 24,186 of 52,038 tokens, 3 s so far
[strata] reading the prompt: 32,378 of 52,038 tokens, 6 s so far
[strata] reading the prompt: 40,570 of 52,038 tokens, 9 s so far
[strata] reading the prompt: 48,762 of 52,038 tokens, 12 s so far
[strata] reading the prompt: 52,031 of 52,038 tokens, 15 s so far
[strata] writing a tool call: execute_code: 48 of max 65536 tokens, 111.8 tok/s, 16 s
[strata] writing a tool call: execute_code: 180 of max 65536 tokens, 125.3 tok/s, 17 s
[strata] writing a tool call: execute_code: 330 of max 65536 tokens, 134.9 tok/s, 18 s
[strata] writing a tool call: execute_code: 478 of max 65536 tokens, 138.1 tok/s, 19 s
[strata] writing a tool call: execute_code: 655 of max 65536 tokens, 146.7 tok/s, 20 s
[strata] writing a tool call: execute_code: 807 of max 65536 tokens, 147.1 tok/s, 21 s
[strata] done: 846 tokens in 22 s (146.3 tok/s) (stop, cancel=False), expert cache 84.5% hit
[strata] reading the prompt: 1,030 of 1,037 tokens, 3 s so far
[strata] done: 4 tokens in 3 s (57.1 tok/s) (stop, cancel=False), expert cache 60.8% hit
[strata] reading the prompt: 8,192 of 53,028 tokens, 3 s so far
[strata] reading the prompt: 15,994 of 53,028 tokens, 6 s so far
[strata] reading the prompt: 24,186 of 53,028 tokens, 9 s so far
[strata] reading the prompt: 32,378 of 53,028 tokens, 12 s so far
[strata] reading the prompt: 40,570 of 53,028 tokens, 15 s so far
[strata] reading the prompt: 48,762 of 53,028 tokens, 18 s so far
[strata] reading the prompt: 53,021 of 53,028 tokens, 21 s so far
[strata] writing a tool call: execute_code: 60 of max 65536 tokens, 135.5 tok/s, 22 s
[strata] writing a tool call: execute_code: 235 of max 65536 tokens, 162.1 tok/s, 23 s
[strata] writing a tool call: execute_code: 428 of max 65536 tokens, 174.2 tok/s, 24 s
[strata] writing a tool call: execute_code: 622 of max 65536 tokens, 179.3 tok/s, 25 s
[strata] writing a tool call: execute_code: 795 of max 65536 tokens, 177.8 tok/s, 26 s
[strata] done: 852 tokens in 26 s (176.2 tok/s) (stop, cancel=False), expert cache 90.0% hit
[strata] reading the prompt: 1,036 of 1,043 tokens, 3 s so far
[strata] done: 4 tokens in 3 s (53.6 tok/s) (stop, cancel=False), expert cache 60.1% hit
[strata] reading the prompt: 8,192 of 54,024 tokens, 3 s so far
[strata] reading the prompt: 15,994 of 54,024 tokens, 6 s so far
[strata] reading the prompt: 24,186 of 54,024 tokens, 9 s so far
[strata] reading the prompt: 32,378 of 54,024 tokens, 12 s so far
[strata] reading the prompt: 40,570 of 54,024 tokens, 15 s so far
[strata] reading the prompt: 48,762 of 54,024 tokens, 18 s so far
[strata] reading the prompt: 54,017 of 54,024 tokens, 21 s so far

[tester12340]

[tester12340]

无论怎么折腾，遇到长下文直接崩掉。。从来。
大神们，我还有啥能做的吗？

[perter]

@tester12340
显存不够还想要多线程的话，还可以把KV缓存降低到q4_0

"--max-context",
"131072",
"--kv",
"q4_0",
"--kv-resident",
"32768",
"--vram-reserve-mib",
"700",

[tester12340]

感谢，测试了。我感觉是可以了。 我再测试一下。
两个并行，单个能跑到 200 tok每秒
writing a tool call: execute_code: 816 of max 65536 tokens, 215.0 tok/s, 6 s
[strata] done: 944 tokens in 6 s (216.2 tok/s) (stop, cancel=False), expert cache 97.5% hit

感谢peter！

[strata] reading the prompt: 47,058 of 47,065 tokens, 2 s so far
[strata] answering: 137 of max 65536 tokens, 176.2 tok/s, 3 s
[strata] answering: 322 of max 65536 tokens, 180.1 tok/s, 4 s
[strata] writing a tool call: execute_code: 560 of max 65536 tokens, 200.3 tok/s, 5 s
[strata] writing a tool call: execute_code: 783 of max 65536 tokens, 205.9 tok/s, 6 s

stara 现在这么牛逼 感觉那些小主机全没用了啊

[tester12340]

为论坛做贡献，分享一下 5090 目前配置 stara 跑 next 的配置。 感谢 Peter 和 各位大神的指点。感谢锤子提供这么好的论坛。

{
"exe": "C:\strata\Strata\engine\strata.exe",
"args": [
"--pack",
"C:\strata\Strata-data\packs\q2_0",
"--native",
"C:\strata\Strata-data\models\Q2_0\Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00001-of-00002.gguf",
"--ple-gguf",
"C:\strata\Strata-data\models\Q2_0\Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00002-of-00002.gguf",
"--expert-profile",
"C:\strata\Strata\data\expert-profile.bin",
"--expert-cache",
"auto",
"--prefill",
"auto",
"--spec",
"4",
"--mtp",
"C:\strata\Strata-data\mtp\rt",
"--max-context",
"131072",
"--kv","q4_0",
"--kv-resident","32768",
"--vram-reserve-mib","700",
"--resident-experts",
"--pcie-frac",
"0.35",
"--spec-min-p",
"0.50",
"--pool-workers",
"4",

"--conversation-cache-mib", "8192"
],
"cwd": "C:\strata\Strata",
"parallel": 4,
"tokenizer": "C:\strata\Strata-data\packs\q2_0\tokenizer",
"model_name": "qwen3.8-flash-next-q2_0",
"log": "C:\strata\Strata\strata-q2_0.log",
"lib_dirs": [
"C:\strata\Strata\.venv\Lib\site-packages\nvidia\cu13\bin\x86_64"
],
"host":"192.168.0.3",
"port": 8080,
"gpu": 0,
"gpus_asked": true
}

有需要优化的地方 请大家指正。

另外默默问一句， mac ultra 96 是不是会完败 5090 staraa?

[David Chen]

我有沒有看錯！我3060那辣雞都跑iq3_s了，你5090只跑q2?

[tester12340]

我需要长上下文啊 兄弟。而且 "--kv","q4_0" 感觉要不要回到 27b了。

[stxpnet]

你真的别用Q2，可能造成工具链断裂，然后无限循环

[tester12340]

我正在体验无限循环。 但我切到 qwen flash 在线也是循环。估计人品不行。

[terry]

@tester12340 你这发帖格式是给论坛做贡献吗？让AI整理成markdwon代码再发

[tester12340]

按terry 老板要求

{
"exe": "C:\strata\Strata\engine\strata.exe",
"args": [
"--pack",
"C:\strata\Strata-data\packs\q2_0",
"--native",
"C:\strata\Strata-data\models\Q2_0\Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00001-of-00002.gguf",
"--ple-gguf",
"C:\strata\Strata-data\models\Q2_0\Qwen3.8-Flash-Next-GSQ-RCO-Q2_0-00002-of-00002.gguf",
"--expert-profile",
"C:\strata\Strata\data\expert-profile.bin",
"--expert-cache",
"auto",
"--prefill",
"auto",
"--spec",
"4",
"--mtp",
"C:\strata\Strata-data\mtp\rt",
"--max-context",
"131072",
"--kv",
"q4_0",
"--kv-resident",
"32768",
"--vram-reserve-mib",
"700",
"--resident-experts",
"--pcie-frac",
"0.35",
"--spec-min-p",
"0.50",
"--pool-workers",
"4",
"--conversation-cache-mib",
"8192"
],
"cwd": "C:\strata\Strata",
"parallel": 4,
"tokenizer": "C:\strata\Strata-data\packs\q2_0\tokenizer",
"model_name": "qwen3.8-flash-next-q2_0",
"log": "C:\strata\Strata\strata-q2_0.log",
"lib_dirs": [
"C:\strata\Strata\.venv\Lib\site-packages\nvidia\cu13\bin\x86_64"
],
"host": "192.168.0.3",
"port": 8080,
"gpu": 0,
"gpus_asked": true
}

这个对了吗？

[tester12340]

@stxpnet 我试试 Q3 能不能跑

[tester12340]

下载中

[tester12340]

这个是不是太慢了？

[tester12340]

从 300tk 直接降到 20tk

[tester12340]

晕倒，我再 4080 上跑了一下，竟然跟 5090 速度差不多 T_T 我了个晕倒。

[Tony Wang]

@tester12340

你用```分别在参数配置部分的上方和下方, 这个表示中间是代码的意思, 也可以一键复制.

建议学一下Mark Down 语法, 比较简单, 常用的也就那么几条.