Compare commits

..

2 Commits

Author SHA1 Message Date
778b04a80f Reapply "llama-cpp: maybe use vulkan?"
All checks were successful
Build and Deploy / deploy (push) Successful in 2m17s
This reverts commit 9addb1569a.
2026-04-07 19:12:57 -04:00
88fc219f2d update 2026-04-07 19:11:50 -04:00
3 changed files with 17 additions and 16 deletions

24
flake.lock generated
View File

@@ -325,16 +325,16 @@
] ]
}, },
"locked": { "locked": {
"lastModified": 1775236905, "lastModified": 1774922513,
"narHash": "sha256-tHshzR/k6D/r5UhJCfJ9b/mJgsbn7ODtnZrDlimhOOI=", "narHash": "sha256-TKk1i8AZzxy4/z0MkqKxoGf/CQDvoL+jo8JDtZeCRy8=",
"owner": "TheTom", "owner": "apollosenvy",
"repo": "llama-cpp-turboquant", "repo": "llama-cpp-turboquant",
"rev": "bc05a6803e48f17e0f2c7a99fce9b50d03882de7", "rev": "9e80e93ceb115bc5055997c373d8c09bfa47a565",
"type": "github" "type": "github"
}, },
"original": { "original": {
"owner": "TheTom", "owner": "apollosenvy",
"ref": "feature/turboquant-kv-cache", "ref": "pr/vulkan-turbo3",
"repo": "llama-cpp-turboquant", "repo": "llama-cpp-turboquant",
"type": "github" "type": "github"
} }
@@ -368,11 +368,11 @@
"systems": "systems_3" "systems": "systems_3"
}, },
"locked": { "locked": {
"lastModified": 1775446111, "lastModified": 1775531897,
"narHash": "sha256-3W1RFYoJgpC9N7Oezj3r4ILOzBP4LSob8QZV0/vuxhc=", "narHash": "sha256-3NIpnV1HxBCwi00iMvj9KcqXkM0VNA72KABj8g0cFFs=",
"owner": "Infinidoge", "owner": "Infinidoge",
"repo": "nix-minecraft", "repo": "nix-minecraft",
"rev": "059dc0e19a275112ba0a396f0d7d2c4cda062d10", "rev": "8c7693880cb861e60adeab5480f02dc3e7a390f6",
"type": "github" "type": "github"
}, },
"original": { "original": {
@@ -715,11 +715,11 @@
"trackerlist": { "trackerlist": {
"flake": false, "flake": false,
"locked": { "locked": {
"lastModified": 1775426970, "lastModified": 1775599784,
"narHash": "sha256-MXs6xRTFxCvXnhShHMTCSw70nFeIkY1L20YWXso0xyo=", "narHash": "sha256-ZapxbiFEYjJV2nhdowHQ/8+c8Jd5fpBIEKDiPEmyNgI=",
"owner": "ngosang", "owner": "ngosang",
"repo": "trackerslist", "repo": "trackerslist",
"rev": "00634b20e7c805cffcde71f280324ef6ab45607f", "rev": "6cc71b5b65349081bb713719f5142c200438a327",
"type": "github" "type": "github"
}, },
"original": { "original": {

View File

@@ -29,7 +29,8 @@
}; };
llamacpp = { llamacpp = {
url = "github:TheTom/llama-cpp-turboquant/feature/turboquant-kv-cache"; # url = "github:TheTom/llama-cpp-turboquant/feature/turboquant-kv-cache";
url = "github:apollosenvy/llama-cpp-turboquant/pr/vulkan-turbo3";
inputs.nixpkgs.follows = "nixpkgs"; inputs.nixpkgs.follows = "nixpkgs";
}; };

View File

@@ -23,10 +23,10 @@ in
); );
port = service_configs.ports.private.llama_cpp.port; port = service_configs.ports.private.llama_cpp.port;
host = "0.0.0.0"; host = "0.0.0.0";
package = (lib.optimizePackage inputs.llamacpp.packages.${pkgs.system}.default); package = (lib.optimizePackage inputs.llamacpp.packages.${pkgs.system}.vulkan);
extraFlags = [ extraFlags = [
# "-ngl" "-ngl"
# "12" "999"
"-c" "-c"
"65536" "65536"
"-ctk" "-ctk"