From 1d7c13638cb96114c31698cbbda74456c8dbed0d Mon Sep 17 00:00:00 2001 From: ChiruGuru99 Date: Sun, 26 Jul 2026 02:54:29 +0530 Subject: [PATCH 1/3] Submit two-hart RMS norm Llama optimization --- ported_models/llama_cpp_et/src/llama.cpp-et | 2 +- .../llama_cpp_et/submissions/llama32_1b.track.json | 7 +++++++ 2 files changed, 8 insertions(+), 1 deletion(-) create mode 100644 ported_models/llama_cpp_et/submissions/llama32_1b.track.json diff --git a/ported_models/llama_cpp_et/src/llama.cpp-et b/ported_models/llama_cpp_et/src/llama.cpp-et index f67c2b2a..26b10115 160000 --- a/ported_models/llama_cpp_et/src/llama.cpp-et +++ b/ported_models/llama_cpp_et/src/llama.cpp-et @@ -1 +1 @@ -Subproject commit f67c2b2a30a24bf9354857cde0f233620a20eb23 +Subproject commit 26b1011518a33e15050796630e829c315ab6704f diff --git a/ported_models/llama_cpp_et/submissions/llama32_1b.track.json b/ported_models/llama_cpp_et/submissions/llama32_1b.track.json new file mode 100644 index 00000000..cfeca0cb --- /dev/null +++ b/ported_models/llama_cpp_et/submissions/llama32_1b.track.json @@ -0,0 +1,7 @@ +{ + "schema_version": 1, + "track": "llama_3_2_1b_fastest", + "model": "llama32_1b", + "submission_id": "chiruguru99-llama32-rms-minion2-v1", + "runtime_revision": "26b1011518a33e15050796630e829c315ab6704f" +} From 960c956c41cd58ec6b3409e887eede619ee2fddd Mon Sep 17 00:00:00 2001 From: ChiruGuru99 Date: Sun, 26 Jul 2026 05:24:27 +0530 Subject: [PATCH 2/3] Add fused RoPE cache update optimization --- ported_models/llama_cpp_et/src/llama.cpp-et | 2 +- ported_models/llama_cpp_et/submissions/llama32_1b.track.json | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/ported_models/llama_cpp_et/src/llama.cpp-et b/ported_models/llama_cpp_et/src/llama.cpp-et index 26b10115..1c58f0cd 160000 --- a/ported_models/llama_cpp_et/src/llama.cpp-et +++ b/ported_models/llama_cpp_et/src/llama.cpp-et @@ -1 +1 @@ -Subproject commit 26b1011518a33e15050796630e829c315ab6704f +Subproject commit 1c58f0cd97a2665e26635a4c4ab0c8e2fafb6ba4 diff --git a/ported_models/llama_cpp_et/submissions/llama32_1b.track.json b/ported_models/llama_cpp_et/submissions/llama32_1b.track.json index cfeca0cb..d7f9589e 100644 --- a/ported_models/llama_cpp_et/submissions/llama32_1b.track.json +++ b/ported_models/llama_cpp_et/submissions/llama32_1b.track.json @@ -2,6 +2,6 @@ "schema_version": 1, "track": "llama_3_2_1b_fastest", "model": "llama32_1b", - "submission_id": "chiruguru99-llama32-rms-minion2-v1", - "runtime_revision": "26b1011518a33e15050796630e829c315ab6704f" + "submission_id": "chiruguru99-llama32-rope-set-fuse-v1", + "runtime_revision": "1c58f0cd97a2665e26635a4c4ab0c8e2fafb6ba4" } From aae3e8999bc1fc19497fdd83a9435d6e4e0b7869 Mon Sep 17 00:00:00 2001 From: ChiruGuru99 Date: Sun, 26 Jul 2026 14:04:17 +0530 Subject: [PATCH 3/3] Add fused residual RMS optimization --- ported_models/llama_cpp_et/src/llama.cpp-et | 2 +- ported_models/llama_cpp_et/submissions/llama32_1b.track.json | 4 ++-- 2 files changed, 3 insertions(+), 3 deletions(-) diff --git a/ported_models/llama_cpp_et/src/llama.cpp-et b/ported_models/llama_cpp_et/src/llama.cpp-et index 1c58f0cd..ec53d562 160000 --- a/ported_models/llama_cpp_et/src/llama.cpp-et +++ b/ported_models/llama_cpp_et/src/llama.cpp-et @@ -1 +1 @@ -Subproject commit 1c58f0cd97a2665e26635a4c4ab0c8e2fafb6ba4 +Subproject commit ec53d562050cde3aeafa2f7330216dbfd52e967f diff --git a/ported_models/llama_cpp_et/submissions/llama32_1b.track.json b/ported_models/llama_cpp_et/submissions/llama32_1b.track.json index d7f9589e..0f0eb6a0 100644 --- a/ported_models/llama_cpp_et/submissions/llama32_1b.track.json +++ b/ported_models/llama_cpp_et/submissions/llama32_1b.track.json @@ -2,6 +2,6 @@ "schema_version": 1, "track": "llama_3_2_1b_fastest", "model": "llama32_1b", - "submission_id": "chiruguru99-llama32-rope-set-fuse-v1", - "runtime_revision": "1c58f0cd97a2665e26635a4c4ab0c8e2fafb6ba4" + "submission_id": "chiruguru99-llama32-add-rms-fuse-v1", + "runtime_revision": "ec53d562050cde3aeafa2f7330216dbfd52e967f" }