Implementation: Ultra-lightweight CPU cycle profiling using RDTSC instruction (~10 cycles overhead). Changes: 1. Added rdtsc() inline function for x86_64 CPU cycle counter 2. Instrumented tiny_fast_alloc(), tiny_fast_free(), tiny_fast_refill() 3. Track malloc, free, refill, and migration cycles separately 4. Profile output via HAKMEM_TINY_PROFILE=1 environment variable 5. Renamed variables to avoid conflict with core/hakmem.c globals Files modified: - core/tiny_fastcache.h: rdtsc(), profile helpers, extern declarations - core/tiny_fastcache.c: counter definitions, print_profile() output Usage: ```bash HAKMEM_TINY_PROFILE=1 ./larson_hakmem 2 8 128 1024 1 12345 4 ``` Results (Larson 4 threads, 1.637M ops/s): ``` [MALLOC] count=20,480, avg_cycles=2,476 [REFILL] count=1,285, avg_cycles=38,412 ← 15.5x slower! [FREE] (no data - not called via fast path) ``` Critical discoveries: 1. **REFILL is the bottleneck:** - Average 38,412 cycles per refill (15.5x slower than malloc) - Refill accounts for: 1,285 × 38,412 = 49.3M cycles - Despite Phase 3 batch optimization, still extremely slow - Calling hak_tiny_alloc() 16 times has massive overhead 2. **MALLOC is 24x slower than expected:** - Average 2,476 cycles (expected ~100 cycles for tcache) - Even cache hits are slow - Profiling overhead is only ~10 cycles, so real cost is ~2,466 cycles - Something fundamentally wrong with fast path 3. **Only 2.5% of allocations use fast path:** - Total operations: 1.637M × 2s = 3.27M ops - Tiny fast alloc: 20,480 × 4 threads = 81,920 ops - Coverage: 81,920 / 3,270,000 = **2.5%** - **97.5% of allocations bypass tiny_fast_alloc entirely!** 4. **FREE is not instrumented:** - No free() calls captured by profiling - hakmem.c's free() likely takes different path - Not calling tiny_fast_free() at all Root cause analysis: The 4x performance gap (vs system malloc) is NOT due to: - Entry point overhead (Phase 1) ❌ - Dual free lists (Phase 2) ❌ - Batch refill efficiency (Phase 3) ❌ The REAL problems: 1. **Tiny fast path is barely used** (2.5% coverage) 2. **Refill is catastrophically slow** (38K cycles) 3. **Even cache hits are 24x too slow** (2.5K cycles) 4. **Free path is completely bypassed** Why system malloc is 4x faster: - System tcache has ~100 cycle malloc - System tcache has ~90% hit rate (vs our 2.5% usage) - System malloc/free are symmetric (we only optimize malloc) Next steps: 1. Investigate why 97.5% bypass tiny_fast_alloc 2. Profile the slow path (hak_alloc_at) that handles 97.5% 3. Understand why even cache hits take 2,476 cycles 4. Instrument free() path to see where frees go 5. May need to optimize slow path instead of fast path This profiling reveals we've been optimizing the wrong thing. The "fast path" is neither fast (2.5K cycles) nor used (2.5%).
233 lines
8.1 KiB
C
233 lines
8.1 KiB
C
// tiny_fastcache.h - Ultra-Simple Tiny Fast Path (System tcache style)
|
|
// Phase 6-3: Bypass Magazine/SuperSlab for Tiny allocations (<=128B)
|
|
// Goal: 3-4 instruction fast path, 70-80% of System tcache performance
|
|
#pragma once
|
|
|
|
#include <stdint.h>
|
|
#include <stddef.h>
|
|
#include <string.h>
|
|
#include <stdlib.h> // For getenv()
|
|
|
|
// ========== Configuration ==========
|
|
|
|
// Enable Tiny Fast Path (default: ON for Phase 6-3)
|
|
#ifndef HAKMEM_TINY_FAST_PATH
|
|
#define HAKMEM_TINY_FAST_PATH 1
|
|
#endif
|
|
|
|
// Tiny class count (sizes: 16, 24, 32, 40, 48, 56, 64, 80, 96, 112, 128)
|
|
#define TINY_FAST_CLASS_COUNT 16
|
|
|
|
// Fast cache capacity per class (default: 64 slots, like System tcache)
|
|
#ifndef TINY_FAST_CACHE_CAP
|
|
#define TINY_FAST_CACHE_CAP 64
|
|
#endif
|
|
|
|
// Tiny size threshold (<=128B goes to fast path)
|
|
#define TINY_FAST_THRESHOLD 128
|
|
|
|
// ========== TLS Cache (System tcache style) ==========
|
|
|
|
// Per-thread fast cache: array of freelist heads (defined in tiny_fastcache.c)
|
|
extern __thread void* g_tiny_fast_cache[TINY_FAST_CLASS_COUNT];
|
|
|
|
// Per-thread cache counts (for capacity management)
|
|
extern __thread uint32_t g_tiny_fast_count[TINY_FAST_CLASS_COUNT];
|
|
|
|
// Initialized flag
|
|
extern __thread int g_tiny_fast_initialized;
|
|
|
|
// ========== Phase 6-7: Dual Free Lists (Phase 2) ==========
|
|
// Separate free staging area to reduce cache line bouncing
|
|
|
|
extern __thread void* g_tiny_fast_free_head[TINY_FAST_CLASS_COUNT];
|
|
extern __thread uint32_t g_tiny_fast_free_count[TINY_FAST_CLASS_COUNT];
|
|
|
|
// ========== RDTSC Profiling (Phase 6-8) ==========
|
|
// Extern declarations for inline functions to access profiling counters
|
|
|
|
extern __thread uint64_t g_tiny_malloc_count;
|
|
extern __thread uint64_t g_tiny_malloc_cycles;
|
|
extern __thread uint64_t g_tiny_free_count;
|
|
extern __thread uint64_t g_tiny_free_cycles;
|
|
extern __thread uint64_t g_tiny_refill_cycles;
|
|
extern __thread uint64_t g_tiny_migration_count;
|
|
extern __thread uint64_t g_tiny_migration_cycles;
|
|
|
|
#ifdef __x86_64__
|
|
static inline uint64_t tiny_fast_rdtsc(void) {
|
|
unsigned int lo, hi;
|
|
__asm__ __volatile__ ("rdtsc" : "=a" (lo), "=d" (hi));
|
|
return ((uint64_t)hi << 32) | lo;
|
|
}
|
|
#else
|
|
static inline uint64_t tiny_fast_rdtsc(void) { return 0; }
|
|
#endif
|
|
|
|
extern int g_profile_enabled;
|
|
static inline int tiny_fast_profile_enabled(void) {
|
|
extern int g_profile_enabled;
|
|
if (__builtin_expect(g_profile_enabled == -1, 0)) {
|
|
const char* env = getenv("HAKMEM_TINY_PROFILE");
|
|
g_profile_enabled = (env && *env && *env != '0') ? 1 : 0;
|
|
}
|
|
return g_profile_enabled;
|
|
}
|
|
|
|
// ========== Size to Class Mapping ==========
|
|
// Inline size-to-class for fast path (O(1) lookup table)
|
|
|
|
static inline int tiny_fast_size_to_class(size_t size) {
|
|
// Optimized: Lookup table for O(1) mapping (vs 11-branch linear search)
|
|
// Table indexed by (size >> 3) for sizes 0-128
|
|
// Class mapping: 0:16B, 1:24B, 2:32B, 3:40B, 4:48B, 5:56B, 6:64B, 7:80B, 8:96B, 9:112B, 10:128B
|
|
|
|
static const int8_t size_to_class_lut[17] = {
|
|
0, // 0-7 → 16B (class 0)
|
|
0, // 8-15 → 16B (class 0)
|
|
0, // 16 → 16B (class 0)
|
|
1, // 17-23 → 24B (class 1)
|
|
1, // 24 → 24B (class 1)
|
|
2, // 25-31 → 32B (class 2)
|
|
2, // 32 → 32B (class 2)
|
|
3, // 33-39 → 40B (class 3)
|
|
3, // 40 → 40B (class 3)
|
|
4, // 41-47 → 48B (class 4)
|
|
4, // 48 → 48B (class 4)
|
|
5, // 49-55 → 56B (class 5)
|
|
5, // 56 → 56B (class 5)
|
|
6, // 57-63 → 64B (class 6)
|
|
6, // 64 → 64B (class 6)
|
|
7, // 65-79 → 80B (class 7)
|
|
8 // 80-95 → 96B (class 8)
|
|
};
|
|
|
|
if (__builtin_expect(size > 128, 0)) return -1; // Not tiny
|
|
|
|
// Fast path: Direct lookup (1-2 instructions!)
|
|
unsigned int idx = size >> 3; // size / 8
|
|
if (__builtin_expect(idx < 17, 1)) {
|
|
return size_to_class_lut[idx];
|
|
}
|
|
|
|
// Size 96-128: class 9-10
|
|
if (size <= 112) return 9; // 112B (class 9)
|
|
return 10; // 128B (class 10)
|
|
}
|
|
|
|
// ========== Forward Declarations ==========
|
|
// Slow path: refill from Magazine/SuperSlab (implemented in tiny_fastcache.c)
|
|
void* tiny_fast_refill(int class_idx);
|
|
void tiny_fast_drain(int class_idx);
|
|
|
|
// ========== Fast Path: Alloc (3-4 instructions!) ==========
|
|
|
|
static inline void* tiny_fast_alloc(size_t size) {
|
|
uint64_t start = tiny_fast_profile_enabled() ? tiny_fast_rdtsc() : 0;
|
|
|
|
// Step 1: Size to class (1-2 instructions, branch predictor friendly)
|
|
int cls = tiny_fast_size_to_class(size);
|
|
if (__builtin_expect(cls < 0, 0)) return NULL; // Not tiny (rare)
|
|
|
|
// Step 2: Pop from alloc_head (hot allocation path)
|
|
void* ptr = g_tiny_fast_cache[cls];
|
|
if (__builtin_expect(ptr != NULL, 1)) {
|
|
// Fast path: Pop head, decrement count
|
|
g_tiny_fast_cache[cls] = *(void**)ptr;
|
|
g_tiny_fast_count[cls]--;
|
|
|
|
if (start) {
|
|
g_tiny_malloc_cycles += (tiny_fast_rdtsc() - start);
|
|
g_tiny_malloc_count++;
|
|
}
|
|
return ptr;
|
|
}
|
|
|
|
// ========================================================================
|
|
// Phase 6-7: Step 2.5: Lazy Migration from free_head (Phase 2)
|
|
// If alloc_head empty but free_head has blocks, migrate with pointer swap
|
|
// This is mimalloc's key optimization: batched migration, zero overhead
|
|
// ========================================================================
|
|
if (__builtin_expect(g_tiny_fast_free_head[cls] != NULL, 0)) {
|
|
uint64_t mig_start = start ? tiny_fast_rdtsc() : 0;
|
|
|
|
// Migrate entire free_head → alloc_head (pointer swap, instant!)
|
|
g_tiny_fast_cache[cls] = g_tiny_fast_free_head[cls];
|
|
g_tiny_fast_count[cls] = g_tiny_fast_free_count[cls];
|
|
g_tiny_fast_free_head[cls] = NULL;
|
|
g_tiny_fast_free_count[cls] = 0;
|
|
|
|
// Now pop one from newly migrated list
|
|
ptr = g_tiny_fast_cache[cls];
|
|
g_tiny_fast_cache[cls] = *(void**)ptr;
|
|
g_tiny_fast_count[cls]--;
|
|
|
|
if (mig_start) {
|
|
g_tiny_migration_cycles += (tiny_fast_rdtsc() - mig_start);
|
|
g_tiny_migration_count++;
|
|
}
|
|
|
|
if (start) {
|
|
g_tiny_malloc_cycles += (tiny_fast_rdtsc() - start);
|
|
g_tiny_malloc_count++;
|
|
}
|
|
return ptr;
|
|
}
|
|
|
|
// Step 3: Slow path - refill from Magazine/SuperSlab
|
|
ptr = tiny_fast_refill(cls);
|
|
|
|
if (start) {
|
|
g_tiny_malloc_cycles += (tiny_fast_rdtsc() - start);
|
|
g_tiny_malloc_count++;
|
|
}
|
|
return ptr;
|
|
}
|
|
|
|
// ========== Fast Path: Free (2-3 instructions!) ==========
|
|
|
|
static inline void tiny_fast_free(void* ptr, size_t size) {
|
|
uint64_t start = tiny_fast_profile_enabled() ? tiny_fast_rdtsc() : 0;
|
|
|
|
// Step 1: Size to class
|
|
int cls = tiny_fast_size_to_class(size);
|
|
if (__builtin_expect(cls < 0, 0)) return; // Not tiny (error)
|
|
|
|
// ========================================================================
|
|
// Phase 6-7: Push to free_head (Phase 2)
|
|
// Separate free staging area reduces cache line contention with alloc_head
|
|
// mimalloc's key insight: alloc/free touch different cache lines
|
|
// ========================================================================
|
|
|
|
// Step 2: Check free_head capacity
|
|
if (__builtin_expect(g_tiny_fast_free_count[cls] >= TINY_FAST_CACHE_CAP, 0)) {
|
|
// Free cache full - drain to Magazine/SuperSlab
|
|
tiny_fast_drain(cls);
|
|
}
|
|
|
|
// Step 3: Push to free_head (separate cache line from alloc_head!)
|
|
*(void**)ptr = g_tiny_fast_free_head[cls];
|
|
g_tiny_fast_free_head[cls] = ptr;
|
|
g_tiny_fast_free_count[cls]++;
|
|
|
|
if (start) {
|
|
g_tiny_free_cycles += (tiny_fast_rdtsc() - start);
|
|
g_tiny_free_count++;
|
|
}
|
|
}
|
|
|
|
// ========== Initialization ==========
|
|
|
|
static inline void tiny_fast_init(void) {
|
|
if (g_tiny_fast_initialized) return;
|
|
|
|
memset(g_tiny_fast_cache, 0, sizeof(g_tiny_fast_cache));
|
|
memset(g_tiny_fast_count, 0, sizeof(g_tiny_fast_count));
|
|
|
|
// Phase 6-7: Initialize dual free lists (Phase 2)
|
|
memset(g_tiny_fast_free_head, 0, sizeof(g_tiny_fast_free_head));
|
|
memset(g_tiny_fast_free_count, 0, sizeof(g_tiny_fast_free_count));
|
|
|
|
g_tiny_fast_initialized = 1;
|
|
}
|