tinymembench v0.4.9 (simple benchmark for memory throughput and latency) ========================================================================== == Memory bandwidth tests == == == == Note 1: 1MB = 1000000 bytes == == Note 2: Results for 'copy' tests show how many bytes can be == == copied per second (adding together read and writen == == bytes would have provided twice higher numbers) == == Note 3: 2-pass copy means that we are using a small temporary buffer == == to first fetch data into it, and only then write it to the == == destination (source -> L1 cache, L1 cache -> destination) == == Note 4: If sample standard deviation exceeds 0.1%, it is shown in == == brackets == ========================================================================== C copy backwards : 18176.4 MB/s (1.4%) C copy backwards (32 byte blocks) : 18275.4 MB/s (1.2%) C copy backwards (64 byte blocks) : 18181.1 MB/s (1.7%) C copy : 17874.6 MB/s (2.2%) C copy prefetched (32 bytes step) : 17776.5 MB/s (5.0%) C copy prefetched (64 bytes step) : 17985.1 MB/s (2.4%) C 2-pass copy : 14654.3 MB/s (1.8%) C 2-pass copy prefetched (32 bytes step) : 13729.8 MB/s (3.4%) C 2-pass copy prefetched (64 bytes step) : 13880.0 MB/s (4.8%) C fill : 40499.7 MB/s (1.1%) C fill (shuffle within 16 byte blocks) : 39973.0 MB/s (1.1%) C fill (shuffle within 32 byte blocks) : 39971.5 MB/s (1.0%) C fill (shuffle within 64 byte blocks) : 40053.3 MB/s (1.1%) --- standard memcpy : 35113.0 MB/s (1.2%) standard memset : 60303.4 MB/s (1.2%) --- MOVSB copy : 20441.9 MB/s (2.7%) MOVSD copy : 19682.0 MB/s (0.9%) SSE2 copy : 19940.9 MB/s (2.2%) SSE2 nontemporal copy : 31903.0 MB/s (1.4%) SSE2 copy prefetched (32 bytes step) : 19395.7 MB/s (2.0%) SSE2 copy prefetched (64 bytes step) : 19592.3 MB/s (0.6%) SSE2 nontemporal copy prefetched (32 bytes step) : 26818.1 MB/s (2.5%) SSE2 nontemporal copy prefetched (64 bytes step) : 27904.5 MB/s (3.7%) SSE2 2-pass copy : 16135.5 MB/s (1.3%) SSE2 2-pass copy prefetched (32 bytes step) : 15329.8 MB/s (1.9%) SSE2 2-pass copy prefetched (64 bytes step) : 15353.4 MB/s (2.0%) SSE2 2-pass nontemporal copy : 7100.3 MB/s (1.1%) SSE2 fill : 41106.7 MB/s (1.1%) SSE2 nontemporal fill : 59906.6 MB/s (1.4%) ========================================================================== == Memory latency test == == == == Average time is measured for random memory accesses in the buffers == == of different sizes. The larger is the buffer, the more significant == == are relative contributions of TLB, L1/L2 cache misses and SDRAM == == accesses. For extremely large buffer sizes we are expecting to see == == page table walk with several requests to SDRAM for almost every == == memory access (though 64MiB is not nearly large enough to experience == == this effect to its fullest). == == == == Note 1: All the numbers are representing extra time, which needs to == == be added to L1 cache latency. The cycle timings for L1 cache == == latency can be usually found in the processor documentation. == == Note 2: Dual random read means that we are simultaneously performing == == two independent memory accesses at a time. In the case if == == the memory subsystem can't handle multiple outstanding == == requests, dual random read has the same timings as two == == single reads performed one after another. == ========================================================================== block size : single random read / dual random read, [MADV_NOHUGEPAGE] 1024 : 0.0 ns / 0.0 ns 2048 : 0.0 ns / 0.0 ns 4096 : 0.0 ns / 0.0 ns 8192 : 0.0 ns / 0.0 ns 16384 : 0.0 ns / 0.0 ns 32768 : 0.0 ns / 0.0 ns 65536 : 0.5 ns / 0.8 ns 131072 : 1.2 ns / 1.6 ns 262144 : 1.6 ns / 1.9 ns 524288 : 2.1 ns / 2.4 ns 1048576 : 2.8 ns / 3.1 ns 2097152 : 4.1 ns / 4.7 ns 4194304 : 8.4 ns / 10.2 ns 8388608 : 11.7 ns / 13.0 ns 16777216 : 16.4 ns / 18.6 ns 33554432 : 33.0 ns / 42.4 ns 67108864 : 56.9 ns / 72.8 ns block size : single random read / dual random read, [MADV_HUGEPAGE] 1024 : 0.0 ns / 0.0 ns 2048 : 0.0 ns / 0.0 ns 4096 : 0.0 ns / 0.0 ns 8192 : 0.0 ns / 0.0 ns 16384 : 0.0 ns / 0.0 ns 32768 : 0.0 ns / 0.0 ns 65536 : 0.3 ns / 0.5 ns 131072 : 1.0 ns / 1.3 ns 262144 : 1.4 ns / 1.5 ns 524288 : 2.0 ns / 2.2 ns 1048576 : 2.6 ns / 2.8 ns 2097152 : 3.8 ns / 3.7 ns 4194304 : 8.2 ns / 10.0 ns 8388608 : 10.8 ns / 12.5 ns 16777216 : 15.8 ns / 17.0 ns 33554432 : 31.1 ns / 39.7 ns 67108864 : 54.3 ns / 67.3 ns