-
Notifications
You must be signed in to change notification settings - Fork 96
OMAP5 uEVM (OMAP5432)
Matthijs van Duin edited this page Mar 21, 2017
·
3 revisions
OMAP5 uEVM (rev G)
Texas Instruments OMAP5432 es2.0 SoC
dual-core ARM Cortex-A15 r2p2 @ 1.5 GHz
2 MiB unified L2 cache (2048 sets × 16 ways × 64 bytes)
128-bit AXI4 interface @ cpu/2 (750 MHz)
two 32-bit DDR3-1066 memory interfaces
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 : 1210.2 MB/s
C copy backwards (32 byte blocks) : 1203.3 MB/s (0.1%)
C copy backwards (64 byte blocks) : 2303.2 MB/s
C copy : 2391.6 MB/s (0.1%)
C copy prefetched (32 bytes step) : 1370.4 MB/s
C copy prefetched (64 bytes step) : 1374.4 MB/s
C 2-pass copy : 1415.3 MB/s
C 2-pass copy prefetched (32 bytes step) : 1235.8 MB/s
C 2-pass copy prefetched (64 bytes step) : 1237.5 MB/s
C fill : 5815.9 MB/s (0.2%)
C fill (shuffle within 16 byte blocks) : 1614.1 MB/s
C fill (shuffle within 32 byte blocks) : 1613.4 MB/s
C fill (shuffle within 64 byte blocks) : 1607.1 MB/s
---
standard memcpy : 2191.4 MB/s
standard memset : 5793.8 MB/s (0.2%)
---
NEON read : 3451.1 MB/s
NEON read prefetched (32 bytes step) : 4227.0 MB/s
NEON read prefetched (64 bytes step) : 4231.6 MB/s
NEON read 2 data streams : 3548.7 MB/s
NEON read 2 data streams prefetched (32 bytes step) : 4335.2 MB/s
NEON read 2 data streams prefetched (64 bytes step) : 4335.3 MB/s
NEON copy : 2425.9 MB/s
NEON copy prefetched (32 bytes step) : 2489.1 MB/s
NEON copy prefetched (64 bytes step) : 2488.7 MB/s
NEON unrolled copy : 2194.3 MB/s
NEON unrolled copy prefetched (32 bytes step) : 2512.9 MB/s
NEON unrolled copy prefetched (64 bytes step) : 2513.9 MB/s
NEON copy backwards : 1218.0 MB/s
NEON copy backwards prefetched (32 bytes step) : 1367.1 MB/s (0.1%)
NEON copy backwards prefetched (64 bytes step) : 1367.1 MB/s
NEON 2-pass copy : 2082.7 MB/s
NEON 2-pass copy prefetched (32 bytes step) : 2288.3 MB/s
NEON 2-pass copy prefetched (64 bytes step) : 2289.4 MB/s
NEON unrolled 2-pass copy : 1439.5 MB/s (0.3%)
NEON unrolled 2-pass copy prefetched (32 bytes step) : 1814.1 MB/s (0.8%)
NEON unrolled 2-pass copy prefetched (64 bytes step) : 1853.6 MB/s (0.5%)
NEON fill : 5805.3 MB/s (0.3%)
NEON fill backwards : 1615.7 MB/s
VFP copy : 2350.7 MB/s
VFP 2-pass copy : 1377.4 MB/s
ARM fill (STRD) : 5793.7 MB/s (0.2%)
ARM fill (STM with 8 registers) : 5773.3 MB/s
ARM fill (STM with 4 registers) : 5780.1 MB/s
ARM copy prefetched (incr pld) : 2485.3 MB/s
ARM copy prefetched (wrap pld) : 2444.4 MB/s
ARM 2-pass copy prefetched (incr pld) : 1658.0 MB/s
ARM 2-pass copy prefetched (wrap pld) : 1620.9 MB/s
==========================================================================
== 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
1024 : 0.0 ns / 0.0 ns
2048 : 0.0 ns / 0.0 ns
4096 : 0.0 ns / 0.1 ns
8192 : 0.0 ns / 0.1 ns
16384 : 0.0 ns / 0.0 ns
32768 : 0.0 ns / 0.1 ns
65536 : 5.2 ns / 8.2 ns
131072 : 8.0 ns / 11.0 ns
262144 : 11.6 ns / 14.7 ns
524288 : 13.4 ns / 16.9 ns
1048576 : 14.5 ns / 18.1 ns
2097152 : 19.4 ns / 24.8 ns
4194304 : 90.7 ns / 138.6 ns
8388608 : 127.1 ns / 178.9 ns
16777216 : 145.8 ns / 197.3 ns
33554432 : 156.4 ns / 208.8 ns
67108864 : 171.7 ns / 234.6 ns
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 : 1101.8 MB/s
C copy backwards (32 byte blocks) : 1096.8 MB/s (0.3%)
C copy backwards (64 byte blocks) : 2162.1 MB/s
C copy : 2232.3 MB/s
C copy prefetched (32 bytes step) : 1263.7 MB/s (0.2%)
C copy prefetched (64 bytes step) : 1265.4 MB/s
C 2-pass copy : 1244.1 MB/s
C 2-pass copy prefetched (32 bytes step) : 1116.2 MB/s
C 2-pass copy prefetched (64 bytes step) : 1117.0 MB/s
C fill : 3836.8 MB/s
C fill (shuffle within 16 byte blocks) : 1519.5 MB/s (0.2%)
C fill (shuffle within 32 byte blocks) : 1519.1 MB/s
C fill (shuffle within 64 byte blocks) : 1508.6 MB/s (0.1%)
---
standard memcpy : 1932.0 MB/s (0.4%)
standard memset : 3794.5 MB/s (0.3%)
---
NEON read : 2843.5 MB/s
NEON read prefetched (32 bytes step) : 3487.2 MB/s
NEON read prefetched (64 bytes step) : 3499.9 MB/s
NEON read 2 data streams : 2896.9 MB/s
NEON read 2 data streams prefetched (32 bytes step) : 3595.4 MB/s (0.3%)
NEON read 2 data streams prefetched (64 bytes step) : 3597.6 MB/s
NEON copy : 2285.8 MB/s
NEON copy prefetched (32 bytes step) : 2407.3 MB/s
NEON copy prefetched (64 bytes step) : 2407.3 MB/s
NEON unrolled copy : 2039.2 MB/s (0.3%)
NEON unrolled copy prefetched (32 bytes step) : 2403.3 MB/s
NEON unrolled copy prefetched (64 bytes step) : 2417.1 MB/s
NEON copy backwards : 1114.5 MB/s
NEON copy backwards prefetched (32 bytes step) : 1257.1 MB/s
NEON copy backwards prefetched (64 bytes step) : 1257.2 MB/s
NEON 2-pass copy : 1610.4 MB/s
NEON 2-pass copy prefetched (32 bytes step) : 1826.6 MB/s (0.3%)
NEON 2-pass copy prefetched (64 bytes step) : 1831.1 MB/s (0.1%)
NEON unrolled 2-pass copy : 1290.7 MB/s (0.9%)
NEON unrolled 2-pass copy prefetched (32 bytes step) : 1549.9 MB/s (1.0%)
NEON unrolled 2-pass copy prefetched (64 bytes step) : 1599.2 MB/s (1.1%)
NEON fill : 3814.3 MB/s (0.6%)
NEON fill backwards : 1519.1 MB/s (0.2%)
VFP copy : 2148.3 MB/s
VFP 2-pass copy : 1207.5 MB/s
ARM fill (STRD) : 3797.0 MB/s
ARM fill (STM with 8 registers) : 3818.1 MB/s (0.2%)
ARM fill (STM with 4 registers) : 3835.9 MB/s (0.2%)
ARM copy prefetched (incr pld) : 2397.7 MB/s
ARM copy prefetched (wrap pld) : 2286.6 MB/s
ARM 2-pass copy prefetched (incr pld) : 1442.2 MB/s
ARM 2-pass copy prefetched (wrap pld) : 1386.1 MB/s
==========================================================================
== 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
1024 : 0.0 ns / 0.0 ns
2048 : 0.0 ns / 0.1 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 : 7.9 ns / 12.1 ns
131072 : 12.1 ns / 16.2 ns
262144 : 17.4 ns / 21.9 ns
524288 : 20.1 ns / 25.2 ns
1048576 : 21.6 ns / 27.0 ns
2097152 : 27.6 ns / 35.6 ns
4194304 : 109.6 ns / 166.9 ns
8388608 : 151.3 ns / 216.9 ns
16777216 : 172.9 ns / 241.2 ns
33554432 : 185.0 ns / 256.2 ns
67108864 : 205.3 ns / 291.6 ns
Kernel 4.9.140-tegra #1 SMP PREEMPT Wed Mar 13 00:32:22 PDT 2019 aarch64 GNU/Linux Under xorg, no compositor active, no browser or other cpu hogs.
tinymembench v0.4.9 (simple benchmark for memory thr
==========================================================================
== 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 : 2949.7 MB/s (3.8%)
C copy backwards (32 byte blocks) : 3011.8 MB/s
C copy backwards (64 byte blocks) : 3029.2 MB/s
C copy : 3642.2 MB/s (4.1%)
C copy prefetched (32 bytes step) : 3824.4 MB/s (0.3%)
C copy prefetched (64 bytes step) : 3825.3 MB/s (0.4%)
C 2-pass copy : 2726.2 MB/s
C 2-pass copy prefetched (32 bytes step) : 2902.6 MB/s (2.5%)
C 2-pass copy prefetched (64 bytes step) : 2928.3 MB/s (0.3%)
C fill : 8541.0 MB/s (0.2%)
C fill (shuffle within 16 byte blocks) : 8518.5 MB/s (2.1%)
C fill (shuffle within 32 byte blocks) : 8537.1 MB/s (0.1%)
C fill (shuffle within 64 byte blocks) : 8528.7 MB/s (0.2%)
---
standard memcpy : 3558.8 MB/s
standard memset : 8520.2 MB/s
---
NEON LDP/STP copy : 3633.9 MB/s (4.2%)
NEON LDP/STP copy pldl2strm (32 bytes step) : 1451.0 MB/s (0.3%)
NEON LDP/STP copy pldl2strm (64 bytes step) : 1450.9 MB/s (0.5%)
NEON LDP/STP copy pldl1keep (32 bytes step) : 3882.5 MB/s (3.9%)
NEON LDP/STP copy pldl1keep (64 bytes step) : 3884.0 MB/s (0.4%)
NEON LD1/ST1 copy : 3630.8 MB/s (0.3%)
NEON STP fill : 8537.8 MB/s
NEON STNP fill : 8544.9 MB/s (1.7%)
ARM LDP/STP copy : 3635.8 MB/s (0.3%)
ARM STP fill : 8544.8 MB/s (0.1%)
ARM STNP fill : 8549.2 MB/s (1.0%)
==========================================================================
== Framebuffer read tests. ==
== ==
== Many ARM devices use a part of the system memory as the framebuffer, ==
== typically mapped as uncached but with write-combining enabled. ==
== Writes to such framebuffers are quite fast, but reads are much ==
== slower and very sensitive to the alignment and the selection of ==
== CPU instructions which are used for accessing memory. ==
== ==
== Many x86 systems allocate the framebuffer in the GPU memory, ==
== accessible for the CPU via a relatively slow PCI-E bus. Moreover, ==
== PCI-E is asymmetric and handles reads a lot worse than writes. ==
== ==
== If uncached framebuffer reads are reasonably fast (at least 100 MB/s ==
== or preferably >300 MB/s), then using the shadow framebuffer layer ==
== is not necessary in Xorg DDX drivers, resulting in a nice overall ==
== performance improvement. For example, the xf86-video-fbturbo DDX ==
== uses this trick. ==
==========================================================================
NEON LDP/STP copy (from framebuffer) : 766.0 MB/s
NEON LDP/STP 2-pass copy (from framebuffer) : 688.8 MB/s
NEON LD1/ST1 copy (from framebuffer) : 770.6 MB/s (0.1%)
NEON LD1/ST1 2-pass copy (from framebuffer) : 681.3 MB/s (0.3%)
ARM LDP/STP copy (from framebuffer) : 766.1 MB/s
ARM LDP/STP 2-pass copy (from framebuffer) : 689.1 MB/s
==========================================================================
== 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 / 0.1 ns
2048 : 0.0 ns / 0.1 ns
4096 : 0.0 ns / 0.1 ns
8192 : 0.0 ns / 0.1 ns
16384 : 0.1 ns / 0.1 ns
32768 : 1.7 ns / 2.9 ns
65536 : 6.4 ns / 9.5 ns
131072 : 9.6 ns / 12.3 ns
262144 : 13.7 ns / 17.0 ns
524288 : 15.8 ns / 19.7 ns
1048576 : 17.3 ns / 22.1 ns
2097152 : 42.1 ns / 64.2 ns
4194304 : 98.5 ns / 138.1 ns
8388608 : 143.9 ns / 186.3 ns
16777216 : 167.2 ns / 211.2 ns
33554432 : 180.1 ns / 227.1 ns
67108864 : 200.0 ns / 260.2 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 : 6.4 ns / 9.4 ns
131072 : 9.5 ns / 12.2 ns
262144 : 11.2 ns / 13.1 ns
524288 : 12.1 ns / 13.5 ns
1048576 : 12.8 ns / 13.6 ns
2097152 : 27.0 ns / 33.0 ns
4194304 : 90.6 ns / 127.8 ns
8388608 : 123.9 ns / 153.8 ns
16777216 : 139.5 ns / 161.2 ns
33554432 : 147.2 ns / 163.6 ns
67108864 : 154.0 ns / 167.6 ns