test: stacked layout string may carry the titan_contiguous suffix

Simon-Pierre Boucher committed 2 days ago (Nov 1, 2026) parent ffe7dec Browse files

3 changed files +28 −1

modified benchmarks/raw/gateG/energy_laptop/bench.jsonl +4 −0

				`@@ -52,3 +52,7 @@`
52	{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": true, "repeat": -1, "family": "code", "prompt": "Write a bash script that lists the ten largest files under a", "ts": "2026-11-01T01:30:29", "prompt_tokens": 19, "gen_tokens": 256, "ttft_ms": 390.1689579943195, "prompt_tps": 48.74155925723839, "gen_tps": 37.77430336988103, "decode_tps_wall": 37.626785667837176, "itl_ms_p50": 24.885875056497753, "itl_ms_p95": 40.63329100608826, "itl_ms_p99": 71.47024990990758, "itl_ms_mean": 26.57681176457189, "total_s": 7.167255957960151, "peak_mem_gb": 1.784851132, "text": "#!/bin/bash\n\n# Specify the directory\ndir=\"/path/to/directory\"\n\n# Get the size of each file and store it in an array\nfiles=( $(ls -l \"$dir\" \| awk '{print $5, $9}') )\n\n# Sort the array in descending ord", "t_start_epoch": 1793514622.7916439, "t_end_epoch": 1793514629.960074, "t_first_epoch": 1793514623.1818128, "cache_hit_rate": 0.848495410712211, "cache_misses": 5216, "cache_hits": 29212, "stall_s": 2.3131685683038086, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 9759, "installed": 9758, "late": 341, "dup": 21945, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 4.139780909870751, "prefetch": 0.5597494770772755, "ensure": 2.1483880928717554, "build": 0.059471825836226344}, "ssd_bytes": 52995686400, "ssd_read_s": 7.924561083083972}	52	{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": true, "repeat": -1, "family": "code", "prompt": "Write a bash script that lists the ten largest files under a", "ts": "2026-11-01T01:30:29", "prompt_tokens": 19, "gen_tokens": 256, "ttft_ms": 390.1689579943195, "prompt_tps": 48.74155925723839, "gen_tps": 37.77430336988103, "decode_tps_wall": 37.626785667837176, "itl_ms_p50": 24.885875056497753, "itl_ms_p95": 40.63329100608826, "itl_ms_p99": 71.47024990990758, "itl_ms_mean": 26.57681176457189, "total_s": 7.167255957960151, "peak_mem_gb": 1.784851132, "text": "#!/bin/bash\n\n# Specify the directory\ndir=\"/path/to/directory\"\n\n# Get the size of each file and store it in an array\nfiles=( $(ls -l \"$dir\" \| awk '{print $5, $9}') )\n\n# Sort the array in descending ord", "t_start_epoch": 1793514622.7916439, "t_end_epoch": 1793514629.960074, "t_first_epoch": 1793514623.1818128, "cache_hit_rate": 0.848495410712211, "cache_misses": 5216, "cache_hits": 29212, "stall_s": 2.3131685683038086, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 9759, "installed": 9758, "late": 341, "dup": 21945, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 4.139780909870751, "prefetch": 0.5597494770772755, "ensure": 2.1483880928717554, "build": 0.059471825836226344}, "ssd_bytes": 52995686400, "ssd_read_s": 7.924561083083972}
53	{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 0, "family": "general", "prompt": "Describe the main differences between a Mac Studio and a Mac", "ts": "2026-11-01T01:30:38", "prompt_tokens": 18, "gen_tokens": 256, "ttft_ms": 838.4084589779377, "prompt_tps": 21.496361739417928, "gen_tps": 32.63569056171431, "decode_tps_wall": 32.50862027189841, "itl_ms_p50": 27.306250063702464, "itl_ms_p95": 54.80616702698171, "itl_ms_p99": 118.00695804413408, "itl_ms_mean": 30.76107172916333, "total_s": 8.682481749914587, "peak_mem_gb": 1.792476808, "text": "\n\nMac Studio:\n- More powerful hardware options, including multiple GPUs and more RAM.\n- Higher-end CPU options, often with more cores and higher clock speeds.\n- Larger display with a higher resolution", "t_start_epoch": 1793514629.960502, "t_end_epoch": 1793514638.643897, "t_first_epoch": 1793514630.7989104, "cache_hit_rate": 0.8172815196617826, "cache_misses": 6310, "cache_hits": 28224, "stall_s": 3.7183112679049373, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 15840, "installed": 15836, "late": 1217, "dup": 15912, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 4.124537539551966, "prefetch": 1.1234240679768845, "ensure": 3.083217271254398, "build": 0.06652727141045034}, "ssd_bytes": 78387609600, "ssd_read_s": 13.75253405154217}	53	{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 0, "family": "general", "prompt": "Describe the main differences between a Mac Studio and a Mac", "ts": "2026-11-01T01:30:38", "prompt_tokens": 18, "gen_tokens": 256, "ttft_ms": 838.4084589779377, "prompt_tps": 21.496361739417928, "gen_tps": 32.63569056171431, "decode_tps_wall": 32.50862027189841, "itl_ms_p50": 27.306250063702464, "itl_ms_p95": 54.80616702698171, "itl_ms_p99": 118.00695804413408, "itl_ms_mean": 30.76107172916333, "total_s": 8.682481749914587, "peak_mem_gb": 1.792476808, "text": "\n\nMac Studio:\n- More powerful hardware options, including multiple GPUs and more RAM.\n- Higher-end CPU options, often with more cores and higher clock speeds.\n- Larger display with a higher resolution", "t_start_epoch": 1793514629.960502, "t_end_epoch": 1793514638.643897, "t_first_epoch": 1793514630.7989104, "cache_hit_rate": 0.8172815196617826, "cache_misses": 6310, "cache_hits": 28224, "stall_s": 3.7183112679049373, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 15840, "installed": 15836, "late": 1217, "dup": 15912, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 4.124537539551966, "prefetch": 1.1234240679768845, "ensure": 3.083217271254398, "build": 0.06652727141045034}, "ssd_bytes": 78387609600, "ssd_read_s": 13.75253405154217}
54	{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 0, "family": "general", "prompt": "Summarize the history of the Rideau Canal in one paragraph.", "ts": "2026-11-01T01:30:45", "prompt_tokens": 14, "gen_tokens": 256, "ttft_ms": 307.55045800469816, "prompt_tps": 45.69629934715006, "gen_tps": 36.686457870160076, "decode_tps_wall": 36.54326812941098, "itl_ms_p50": 27.61320793069899, "itl_ms_p95": 35.31170799396932, "itl_ms_p99": 41.34791693650186, "itl_ms_mean": 27.36482124309987, "total_s": 7.285579874995165, "peak_mem_gb": 1.753644552, "text": "\n\nThe Rideau Canal, officially opened in 1832, was constructed by the British to connect Ottawa to Montreal via the Ottawa River. It was designed by Colonel John By and took approximately three years ", "t_start_epoch": 1793514638.644236, "t_end_epoch": 1793514645.9310138, "t_first_epoch": 1793514638.9517865, "cache_hit_rate": 0.8069263563627869, "cache_misses": 6612, "cache_hits": 27634, "stall_s": 2.490456422092393, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 16192, "installed": 16197, "late": 740, "dup": 15400, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 3.88708005938679, "prefetch": 0.6125109859276563, "ensure": 2.402447963017039, "build": 0.06990473286714405}, "ssd_bytes": 80702078976, "ssd_read_s": 10.351799859432504}	54	{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 0, "family": "general", "prompt": "Summarize the history of the Rideau Canal in one paragraph.", "ts": "2026-11-01T01:30:45", "prompt_tokens": 14, "gen_tokens": 256, "ttft_ms": 307.55045800469816, "prompt_tps": 45.69629934715006, "gen_tps": 36.686457870160076, "decode_tps_wall": 36.54326812941098, "itl_ms_p50": 27.61320793069899, "itl_ms_p95": 35.31170799396932, "itl_ms_p99": 41.34791693650186, "itl_ms_mean": 27.36482124309987, "total_s": 7.285579874995165, "peak_mem_gb": 1.753644552, "text": "\n\nThe Rideau Canal, officially opened in 1832, was constructed by the British to connect Ottawa to Montreal via the Ottawa River. It was designed by Colonel John By and took approximately three years ", "t_start_epoch": 1793514638.644236, "t_end_epoch": 1793514645.9310138, "t_first_epoch": 1793514638.9517865, "cache_hit_rate": 0.8069263563627869, "cache_misses": 6612, "cache_hits": 27634, "stall_s": 2.490456422092393, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 16192, "installed": 16197, "late": 740, "dup": 15400, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 3.88708005938679, "prefetch": 0.6125109859276563, "ensure": 2.402447963017039, "build": 0.06990473286714405}, "ssd_bytes": 80702078976, "ssd_read_s": 10.351799859432504}
		55	+{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 0, "family": "code", "prompt": "Write a bash script that lists the ten largest files under a", "ts": "2026-11-01T01:30:51", "prompt_tokens": 19, "gen_tokens": 256, "ttft_ms": 321.9717079773545, "prompt_tps": 59.18740368732463, "gen_tps": 50.65934949089925, "decode_tps_wall": 50.46156292168384, "itl_ms_p50": 19.861834007315338, "itl_ms_p95": 24.844833998940885, "itl_ms_p99": 26.26283303834498, "itl_ms_mean": 19.817063564836396, "total_s": 5.375322917010635, "peak_mem_gb": 1.781200948, "text": "#!/bin/bash\n\n# Specify the directory\ndir=\"/path/to/directory\"\n\n# Get the size of each file and store it in an array\nfiles=( $(ls -l \"$dir\" \| awk '{print $5, $9}') )\n\n# Sort the array in descending ord", "t_start_epoch": 1793514645.93139, "t_end_epoch": 1793514651.30745, "t_first_epoch": 1793514646.2533617, "cache_hit_rate": 0.8482920878354827, "cache_misses": 5223, "cache_hits": 29205, "stall_s": 1.4466036556987092, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 9784, "installed": 9784, "late": 125, "dup": 21920, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 3.3612594555597752, "prefetch": 0.28280831524170935, "ensure": 1.4808467627735808, "build": 0.045205815229564905}, "ssd_bytes": 53108932608, "ssd_read_s": 5.679458281258121}
		56	+{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 0, "family": "code", "prompt": "Write a C function that computes the dot product of two floa", "ts": "2026-11-01T01:30:56", "prompt_tokens": 24, "gen_tokens": 256, "ttft_ms": 321.83154206722975, "prompt_tps": 75.0366976534093, "gen_tps": 50.18351287297188, "decode_tps_wall": 49.98757009104001, "itl_ms_p50": 20.094042061828077, "itl_ms_p95": 25.694333016872406, "itl_ms_p99": 27.389249997213483, "itl_ms_mean": 20.00497319991244, "total_s": 5.4230997080449015, "peak_mem_gb": 1.937415632, "text": "#include <arm_neon.h>\n\nfloat dot_product_neon(const float a, const float b, int len) {\n // TODO: Implement the dot product computation using NEON intrinsics\n // Start by loading the arrays int", "t_start_epoch": 1793514651.3075972, "t_end_epoch": 1793514656.731461, "t_first_epoch": 1793514651.6294286, "cache_hit_rate": 0.838486889239146, "cache_misses": 5636, "cache_hits": 29259, "stall_s": 1.4255401014816016, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 10451, "installed": 10451, "late": 112, "dup": 21485, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 3.4393598578171805, "prefetch": 0.2789632437052205, "ensure": 1.4602325750747696, "build": 0.04396384267602116}, "ssd_bytes": 56930992128, "ssd_read_s": 5.710928951739334}
		57	+{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 1, "family": "general", "prompt": "Describe the main differences between a Mac Studio and a Mac", "ts": "2026-11-01T01:31:03", "prompt_tokens": 18, "gen_tokens": 256, "ttft_ms": 280.00575001351535, "prompt_tps": 64.36067721757628, "gen_tps": 40.35307071785748, "decode_tps_wall": 40.195500143293714, "itl_ms_p50": 24.73545796237886, "itl_ms_p95": 28.866332955658436, "itl_ms_p99": 31.44495899323374, "itl_ms_mean": 24.87840669813986, "total_s": 6.6239994580391794, "peak_mem_gb": 1.83841756, "text": "\n\nMac Studio:\n- More powerful hardware options, including multiple GPUs and more RAM.\n- Higher-end CPU options, often with more cores and higher clock speeds.\n- Larger display with a higher resolution", "t_start_epoch": 1793514656.731574, "t_end_epoch": 1793514663.356687, "t_first_epoch": 1793514657.0115798, "cache_hit_rate": 0.8177448311808653, "cache_misses": 6294, "cache_hits": 28240, "stall_s": 1.6402923869900405, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 15938, "installed": 15938, "late": 98, "dup": 15814, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 4.341446908423677, "prefetch": 0.333583117229864, "ensure": 1.6837690656539053, "build": 0.04841878020670265}, "ssd_bytes": 78677803008, "ssd_read_s": 7.641431345953606}
		58	+{"backend": "titan", "capacity": 16, "policy": "lru", "predictor": "probe", "m": 8, "prefetch_workers": 2, "fused_upgate": false, "nocache": true, "max_tokens": 256, "split": "heldout", "tag": "energy", "model": "/Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b", "cmd": "benchmarks/scripts/bench_decode.py --model /Users/simon-pierreboucher/.cache/huggingface/hub/models--mlx-community--OLMoE-1B-7B-0125-Instruct-4bit/snapshots/e23844197887b031e7ddddbb0b8959c5a6853a7b --backend titan --capacity 16 --nocache --predictor probe --m 8 --traces benchmarks/raw/gateC/traces_olmoe_v2_laptop --families general,code --split heldout --repeats 3 --max-tokens 256 --tag energy --out benchmarks/raw/gateG/energy_laptop/bench.jsonl", "env": {"host": "MacBookPro-de-simon-pierre-2.local", "chip": "Apple M5 Max", "ram_gb": 48, "macos": "27.0", "python": "3.12.13", "mlx": "0.32.4", "mlx_lm": "0.32.2", "numpy": "2.5.4", "git": "e0179f6"}, "seed": 0, "warmup": false, "repeat": 1, "family": "code", "prompt": "Write a C function that computes the dot product of two floa", "ts": "2026-11-01T01:31:09", "prompt_tokens": 24, "gen_tokens": 256, "ttft_ms": 317.7586660021916, "prompt_tps": 75.61694756174724, "gen_tps": 44.943455296321034, "decode_tps_wall": 44.767957139778446, "itl_ms_p50": 22.643874981440604, "itl_ms_p95": 26.98620897717774, "itl_ms_p99": 28.52170797996223, "itl_ms_mean": 22.337405231105638, "total_s": 6.013796999934129, "peak_mem_gb": 1.944690112, "text": "#include <arm_neon.h>\n\nfloat dot_product_neon(const float a, const float b, int len) {\n // TODO: Implement the dot product computation using NEON intrinsics\n // Start by loading the arrays int", "t_start_epoch": 1793514663.356802, "t_end_epoch": 1793514669.371687, "t_first_epoch": 1793514663.6745605, "cache_hit_rate": 0.838486889239146, "cache_misses": 5636, "cache_hits": 29259, "stall_s": 1.4071811041794717, "hits_device": 0, "layers_complete": 0, "prefetch": {"issued": 10474, "installed": 10474, "late": 16, "dup": 21462, "predictor": "probe", "m": 8}, "host_time_s": {"sync": 4.0537837136071175, "prefetch": 0.262417831341736, "ensure": 1.4553386754123494, "build": 0.04409997153561562}, "ssd_bytes": 57012387840, "ssd_read_s": 5.6420947627630085}

modified benchmarks/raw/gateG/energy_laptop/power.jsonl +23 −0

+ {"t": 1793514644.979489, "elapsed_ns": 1023149708, "cpu_mW": 24534.0, "gpu_mW": 5829.06, "ane_mW": 0.0, "combined_mW": 30363.1, "thermal_pressure": "Nominal"}
+ {"t": 1793514646.002538, "elapsed_ns": 1022825125, "cpu_mW": 22665.7, "gpu_mW": 1668.91, "ane_mW": 0.0, "combined_mW": 24334.6, "thermal_pressure": "Nominal"}
+ {"t": 1793514647.0245628, "elapsed_ns": 1021656541, "cpu_mW": 26014.6, "gpu_mW": 9672.53, "ane_mW": 0.0, "combined_mW": 35687.1, "thermal_pressure": "Nominal"}
++{"t": 1793514648.0464041, "elapsed_ns": 1023803166, "cpu_mW": 23506.5, "gpu_mW": 12.6978, "ane_mW": 0.0, "combined_mW": 23519.2, "thermal_pressure": "Nominal"}
++{"t": 1793514649.068141, "elapsed_ns": 1022234125, "cpu_mW": 15532.6, "gpu_mW": 15322.3, "ane_mW": 0.0, "combined_mW": 30855.0, "thermal_pressure": "Nominal"}
++{"t": 1793514651.1116061, "elapsed_ns": 1021507416, "cpu_mW": 14704.7, "gpu_mW": 2747.9, "ane_mW": 0.0, "combined_mW": 17452.6, "thermal_pressure": "Nominal"}
++{"t": 1793514651.1135008, "elapsed_ns": 1022153125, "cpu_mW": 15082.9, "gpu_mW": 2180.69, "ane_mW": 0.0, "combined_mW": 17263.6, "thermal_pressure": "Nominal"}
++{"t": 1793514652.132927, "elapsed_ns": 1021306208, "cpu_mW": 15940.4, "gpu_mW": 2221.66, "ane_mW": 0.0, "combined_mW": 18162.0, "thermal_pressure": "Nominal"}
++{"t": 1793514653.154567, "elapsed_ns": 1022127250, "cpu_mW": 16136.0, "gpu_mW": 2514.36, "ane_mW": 0.0, "combined_mW": 18650.3, "thermal_pressure": "Nominal"}
++{"t": 1793514654.176312, "elapsed_ns": 1021609333, "cpu_mW": 16913.5, "gpu_mW": 2362.94, "ane_mW": 0.0, "combined_mW": 19276.4, "thermal_pressure": "Nominal"}
++{"t": 1793514655.198668, "elapsed_ns": 1021520583, "cpu_mW": 16625.2, "gpu_mW": 1976.47, "ane_mW": 0.0, "combined_mW": 18601.7, "thermal_pressure": "Nominal"}
++{"t": 1793514656.220159, "elapsed_ns": 1021733791, "cpu_mW": 15003.9, "gpu_mW": 1866.44, "ane_mW": 0.0, "combined_mW": 16870.3, "thermal_pressure": "Nominal"}
++{"t": 1793514657.241514, "elapsed_ns": 1022104666, "cpu_mW": 16129.5, "gpu_mW": 2148.51, "ane_mW": 0.0, "combined_mW": 18278.0, "thermal_pressure": "Nominal"}
++{"t": 1793514658.262487, "elapsed_ns": 1021120166, "cpu_mW": 16368.3, "gpu_mW": 2260.26, "ane_mW": 0.0, "combined_mW": 18628.6, "thermal_pressure": "Nominal"}
++{"t": 1793514660.306035, "elapsed_ns": 1021108791, "cpu_mW": 15137.5, "gpu_mW": 1501.31, "ane_mW": 0.0, "combined_mW": 16638.8, "thermal_pressure": "Nominal"}
++{"t": 1793514660.3077781, "elapsed_ns": 1021776375, "cpu_mW": 15614.0, "gpu_mW": 1469.01, "ane_mW": 0.0, "combined_mW": 17083.0, "thermal_pressure": "Nominal"}
++{"t": 1793514661.328289, "elapsed_ns": 1021469624, "cpu_mW": 16051.4, "gpu_mW": 1478.26, "ane_mW": 0.0, "combined_mW": 17529.6, "thermal_pressure": "Nominal"}
++{"t": 1793514662.350026, "elapsed_ns": 1021858125, "cpu_mW": 15607.8, "gpu_mW": 1488.46, "ane_mW": 0.0, "combined_mW": 17096.3, "thermal_pressure": "Nominal"}
++{"t": 1793514663.3735511, "elapsed_ns": 1022393166, "cpu_mW": 14615.7, "gpu_mW": 1466.17, "ane_mW": 0.0, "combined_mW": 16081.9, "thermal_pressure": "Nominal"}
++{"t": 1793514664.393475, "elapsed_ns": 1021964375, "cpu_mW": 16110.2, "gpu_mW": 1428.62, "ane_mW": 0.0, "combined_mW": 17538.8, "thermal_pressure": "Nominal"}
++{"t": 1793514665.41536, "elapsed_ns": 1021849958, "cpu_mW": 15581.5, "gpu_mW": 2090.33, "ane_mW": 0.0, "combined_mW": 17671.9, "thermal_pressure": "Nominal"}
++{"t": 1793514666.437129, "elapsed_ns": 1021458208, "cpu_mW": 15358.4, "gpu_mW": 1487.09, "ane_mW": 0.0, "combined_mW": 16845.5, "thermal_pressure": "Nominal"}
++{"t": 1793514668.481022, "elapsed_ns": 1021986666, "cpu_mW": 15189.0, "gpu_mW": 1488.28, "ane_mW": 0.0, "combined_mW": 16677.3, "thermal_pressure": "Nominal"}
++{"t": 1793514668.483244, "elapsed_ns": 1021877041, "cpu_mW": 14014.4, "gpu_mW": 1470.82, "ane_mW": 0.0, "combined_mW": 15485.2, "thermal_pressure": "Nominal"}
++{"t": 1793514669.5031161, "elapsed_ns": 1021871875, "cpu_mW": 14131.9, "gpu_mW": 9.78596, "ane_mW": 0.0, "combined_mW": 14141.7, "thermal_pressure": "Nominal"}
++{"t": 1793514670.525273, "elapsed_ns": 1021980208, "cpu_mW": 13655.8, "gpu_mW": 2924.71, "ane_mW": 0.0, "combined_mW": 16580.6, "thermal_pressure": "Nominal"}

modified tests/test_registry_stacked.py +1 −1

				`@@ -10,7 +10,7 @@ def _dir():`
10	@pytest.mark.skipif(_dir() is None, reason="Qwen3-30B-A3B-Instruct-2507-4bit not cached")	10	@pytest.mark.skipif(_dir() is None, reason="Qwen3-30B-A3B-Instruct-2507-4bit not cached")
11	def test_stacked_slices_match_mx_load():	11	def test_stacked_slices_match_mx_load():
12	d = _dir(); s = load_spec(d)	12	d = _dir(); s = load_spec(d)
13	− assert s.layout == "stacked" and s.num_layers == 48 and s.num_experts == 128 and s.top_k == 8	13	+ assert s.layout.startswith("stacked") and s.num_layers == 48 and s.num_experts == 128 and s.top_k == 8
14	st = ExpertStore(s)	14	st = ExpertStore(s)
15	for (l, e) in [(0, 0), (23, 77), (47, 127)]:	15	for (l, e) in [(0, 0), (23, 77), (47, 127)]:
16	t = st.read_expert(l, e)	16	t = st.read_expert(l, e)
17		17

spb/wp33_uqo

test: stacked layout string may carry the titan_contiguous suffix

Theme