mirror of
https://github.com/dogkeeper886/ollama37.git
synced 2025-12-10 15:57:04 +00:00
ef14fb5b26
This commit represents a complete rework after pulling the latest changes from official ollama/ollama repository and re-applying Tesla K80 compatibility patches. ## Key Changes ### CUDA Compute Capability 3.7 Support (Tesla K80) - Added sm_37 (compute 3.7) to CMAKE_CUDA_ARCHITECTURES in CMakeLists.txt - Updated CMakePresets.json to include compute 3.7 in "CUDA 11" preset - Using 37-virtual (PTX with JIT compilation) for maximum compatibility ### Legacy Toolchain Compatibility - **NVIDIA Driver**: 470.256.02 (last version supporting Kepler/K80) - **CUDA Version**: 11.4.4 (last CUDA 11.x supporting compute 3.7) - **GCC Version**: 10.5.0 (required by CUDA 11.4 host_config.h) ### CPU Architecture Trade-offs Due to GCC 10.5 limitation, sacrificed newer CPU optimizations: - Alderlake CPU variant enabled WITHOUT AVX_VNNI (requires GCC 11+) - Still supports: SSE4.2, AVX, F16C, AVX2, BMI2, FMA - Performance impact: ~3-7% on newer CPUs (acceptable for K80 compatibility) ### Build System Updates - Modified ml/backend/ggml/ggml/src/ggml-cuda/CMakeLists.txt for compute 3.7 - Added -Wno-deprecated-gpu-targets flag to suppress warnings - Updated ml/backend/ggml/ggml/src/CMakeLists.txt for Alderlake without AVX_VNNI ### Upstream Sync Merged latest llama.cpp changes including: - Enhanced KV cache management with ISWA and hybrid memory support - Improved multi-modal support (mtmd framework) - New model architectures (Gemma3, Llama4, Qwen3, etc.) - GPU backend improvements for CUDA, Metal, and ROCm - Updated quantization support and GGUF format handling ### Documentation - Updated CLAUDE.md with comprehensive build instructions - Documented toolchain constraints and CPU architecture trade-offs - Removed outdated CI/CD workflows (tesla-k80-*.yml) - Cleaned up temporary development artifacts ## Rationale This fork maintains Tesla K80 GPU support (compute 3.7) which was dropped in official Ollama due to legacy driver/CUDA requirements. The toolchain constraint creates a deadlock: - K80 → Driver 470 → CUDA 11.4 → GCC 10 → No AVX_VNNI We accept the loss of cutting-edge CPU optimizations to enable running modern LLMs on legacy but still capable Tesla K80 hardware (12GB VRAM per GPU). 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
75 lines
1.9 KiB
Go
75 lines
1.9 KiB
Go
package discover
|
|
|
|
import (
|
|
"log/slog"
|
|
"path/filepath"
|
|
"sort"
|
|
"strings"
|
|
|
|
"github.com/ollama/ollama/format"
|
|
"github.com/ollama/ollama/ml"
|
|
)
|
|
|
|
type memInfo struct {
|
|
TotalMemory uint64 `json:"total_memory,omitempty"`
|
|
FreeMemory uint64 `json:"free_memory,omitempty"`
|
|
FreeSwap uint64 `json:"free_swap,omitempty"` // TODO split this out for system only
|
|
}
|
|
|
|
// CPU type represents a CPU Package occupying a socket
|
|
type CPU struct {
|
|
ID string `cpuinfo:"processor"`
|
|
VendorID string `cpuinfo:"vendor_id"`
|
|
ModelName string `cpuinfo:"model name"`
|
|
CoreCount int
|
|
EfficiencyCoreCount int // Performance = CoreCount - Efficiency
|
|
ThreadCount int
|
|
}
|
|
|
|
func LogDetails(devices []ml.DeviceInfo) {
|
|
sort.Sort(sort.Reverse(ml.ByFreeMemory(devices))) // Report devices in order of scheduling preference
|
|
for _, dev := range devices {
|
|
var libs []string
|
|
for _, dir := range dev.LibraryPath {
|
|
if strings.Contains(dir, filepath.Join("lib", "ollama")) {
|
|
libs = append(libs, filepath.Base(dir))
|
|
}
|
|
}
|
|
typeStr := "discrete"
|
|
if dev.Integrated {
|
|
typeStr = "iGPU"
|
|
}
|
|
slog.Info("inference compute",
|
|
"id", dev.ID,
|
|
"filtered_id", dev.FilteredID,
|
|
"library", dev.Library,
|
|
"compute", dev.Compute(),
|
|
"name", dev.Name,
|
|
"description", dev.Description,
|
|
"libdirs", strings.Join(libs, ","),
|
|
"driver", dev.Driver(),
|
|
"pci_id", dev.PCIID,
|
|
"type", typeStr,
|
|
"total", format.HumanBytes2(dev.TotalMemory),
|
|
"available", format.HumanBytes2(dev.FreeMemory),
|
|
)
|
|
}
|
|
// CPU inference
|
|
if len(devices) == 0 {
|
|
dev, _ := GetCPUMem()
|
|
slog.Info("inference compute",
|
|
"id", "cpu",
|
|
"library", "cpu",
|
|
"compute", "",
|
|
"name", "cpu",
|
|
"description", "cpu",
|
|
"libdirs", "ollama",
|
|
"driver", "",
|
|
"pci_id", "",
|
|
"type", "",
|
|
"total", format.HumanBytes2(dev.TotalMemory),
|
|
"available", format.HumanBytes2(dev.FreeMemory),
|
|
)
|
|
}
|
|
}
|