Skip to content

AMD Ryzen 9 3900X 12 Core Processor

David Huang edited this page Sep 4, 2019 · 3 revisions

AMD Ryzen 9 3900X @ 3.8 GHz (turbo boost disabled)

2×16GB DDR4 3000 MHz 15-17-17-36 (running at 16-17-17-36)
FCLK @ 1500 MHz

$ lscpu
Architecture:        x86_64
CPU op-mode(s):      32-bit, 64-bit
Byte Order:          Little Endian
Address sizes:       48 bits physical, 48 bits virtual
CPU(s):              24
On-line CPU(s) list: 0-23
Thread(s) per core:  2
Core(s) per socket:  12
Socket(s):           1
NUMA node(s):        1
Vendor ID:           AuthenticAMD
CPU family:          23
Model:               113
Model name:          AMD Ryzen 9 3900X 12-Core Processor
Stepping:            0
CPU MHz:             3800.002
BogoMIPS:            7600.00
Hypervisor vendor:   Microsoft
Virtualization type: full
L1d cache:           32K
L1i cache:           32K
L2 cache:            512K
L3 cache:            16384K
NUMA node0 CPU(s):   0-23
Flags:               fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm rep_good nopl cpuid extd_apicid pni pclmulqdq ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand hypervisor lahf_lm cmp_legacy cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw topoext cpb ssbd ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves xsaveerptr virt_ssbd arat rdpid
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                                     :   9698.2 MB/s (0.6%)
 C copy backwards (32 byte blocks)                    :   9739.2 MB/s (0.5%)
 C copy backwards (64 byte blocks)                    :   9720.6 MB/s (0.6%)
 C copy                                               :   9511.3 MB/s (0.4%)
 C copy prefetched (32 bytes step)                    :   9776.4 MB/s (0.4%)
 C copy prefetched (64 bytes step)                    :   9824.3 MB/s (0.6%)
 C 2-pass copy                                        :   8139.8 MB/s (0.8%)
 C 2-pass copy prefetched (32 bytes step)             :   8701.3 MB/s (0.4%)
 C 2-pass copy prefetched (64 bytes step)             :   8661.3 MB/s (0.3%)
 C fill                                               :  13254.4 MB/s (0.8%)
 C fill (shuffle within 16 byte blocks)               :  13269.5 MB/s (0.8%)
 C fill (shuffle within 32 byte blocks)               :  13364.5 MB/s (0.8%)
 C fill (shuffle within 64 byte blocks)               :  13361.8 MB/s (0.9%)
 ---
 standard memcpy                                      :   9563.6 MB/s (0.5%)
 standard memset                                      :  14332.2 MB/s (0.7%)
 ---
 MOVSB copy                                           :   8441.4 MB/s (1.0%)
 MOVSD copy                                           :   8374.1 MB/s (0.6%)
 SSE2 copy                                            :  10402.5 MB/s (0.5%)
 SSE2 nontemporal copy                                :  17751.5 MB/s (0.6%)
 SSE2 copy prefetched (32 bytes step)                 :  10270.2 MB/s (0.5%)
 SSE2 copy prefetched (64 bytes step)                 :  10310.9 MB/s (0.9%)
 SSE2 nontemporal copy prefetched (32 bytes step)     :  18770.4 MB/s (0.5%)
 SSE2 nontemporal copy prefetched (64 bytes step)     :  18815.6 MB/s (0.6%)
 SSE2 2-pass copy                                     :   9385.2 MB/s (0.6%)
 SSE2 2-pass copy prefetched (32 bytes step)          :   9364.1 MB/s (0.5%)
 SSE2 2-pass copy prefetched (64 bytes step)          :   9349.2 MB/s (0.5%)
 SSE2 2-pass nontemporal copy                         :   5313.2 MB/s (0.6%)
 SSE2 fill                                            :  14232.8 MB/s (0.8%)
 SSE2 nontemporal fill                                :  23478.6 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          /     0.0 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 :    1.1 ns          /     1.5 ns
    131072 :    1.6 ns          /     1.9 ns
    262144 :    1.9 ns          /     2.1 ns
    524288 :    3.3 ns          /     3.9 ns
   1048576 :    7.2 ns          /     9.2 ns
   2097152 :    9.3 ns          /    10.8 ns
   4194304 :   10.5 ns          /    11.4 ns
   8388608 :   12.3 ns          /    12.9 ns
  16777216 :   30.3 ns          /    37.4 ns
  33554432 :   67.5 ns          /    89.3 ns
  67108864 :   88.4 ns          /   108.9 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 :    1.1 ns          /     1.5 ns
    131072 :    1.6 ns          /     1.9 ns
    262144 :    1.9 ns          /     2.1 ns
    524288 :    3.2 ns          /     3.9 ns
   1048576 :    7.3 ns          /     9.3 ns
   2097152 :    9.3 ns          /    10.8 ns
   4194304 :   10.2 ns          /    11.3 ns
   8388608 :   11.0 ns          /    11.9 ns
  16777216 :   26.9 ns          /    35.8 ns
  33554432 :   64.4 ns          /    85.1 ns
  67108864 :   81.3 ns          /   100.8 ns

2×16GB DDR4 3000 MHz 15-17-17-36 (running at 16-17-17-36)
FCLK @ 1800 MHz

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                                     :  10296.8 MB/s (0.3%)
 C copy backwards (32 byte blocks)                    :  10274.1 MB/s (0.4%)
 C copy backwards (64 byte blocks)                    :  10302.2 MB/s (0.4%)
 C copy                                               :  10350.9 MB/s (0.4%)
 C copy prefetched (32 bytes step)                    :  10612.2 MB/s (0.3%)
 C copy prefetched (64 bytes step)                    :  10600.9 MB/s (0.2%)
 C 2-pass copy                                        :   8691.7 MB/s (0.3%)
 C 2-pass copy prefetched (32 bytes step)             :   9547.3 MB/s (0.3%)
 C 2-pass copy prefetched (64 bytes step)             :   9531.6 MB/s (0.2%)
 C fill                                               :  13218.0 MB/s (0.3%)
 C fill (shuffle within 16 byte blocks)               :  13109.3 MB/s (0.4%)
 C fill (shuffle within 32 byte blocks)               :  13075.9 MB/s (0.2%)
 C fill (shuffle within 64 byte blocks)               :  13095.8 MB/s (0.5%)
 ---
 standard memcpy                                      :   9556.9 MB/s (0.5%)
 standard memset                                      :  15315.7 MB/s (1.5%)
 ---
 MOVSB copy                                           :   8447.0 MB/s (0.4%)
 MOVSD copy                                           :   8438.5 MB/s (0.5%)
 SSE2 copy                                            :  11176.2 MB/s (0.5%)
 SSE2 nontemporal copy                                :  18370.3 MB/s
 SSE2 copy prefetched (32 bytes step)                 :  10849.6 MB/s (0.2%)
 SSE2 copy prefetched (64 bytes step)                 :  10807.8 MB/s (0.2%)
 SSE2 nontemporal copy prefetched (32 bytes step)     :  19129.7 MB/s
 SSE2 nontemporal copy prefetched (64 bytes step)     :  19161.9 MB/s
 SSE2 2-pass copy                                     :  10255.8 MB/s (0.2%)
 SSE2 2-pass copy prefetched (32 bytes step)          :  10169.9 MB/s (0.2%)
 SSE2 2-pass copy prefetched (64 bytes step)          :  10164.3 MB/s (0.3%)
 SSE2 2-pass nontemporal copy                         :   5570.5 MB/s (0.5%)
 SSE2 fill                                            :  14928.5 MB/s (0.5%)
 SSE2 nontemporal fill                                :  27951.6 MB/s (0.8%)

==========================================================================
== 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 :    1.0 ns          /     1.4 ns
    131072 :    1.5 ns          /     1.8 ns
    262144 :    1.8 ns          /     1.9 ns
    524288 :    2.1 ns          /     2.2 ns
   1048576 :    5.4 ns          /     7.1 ns
   2097152 :    8.0 ns          /     9.5 ns
   4194304 :    9.4 ns          /    10.3 ns
   8388608 :   10.8 ns          /    10.9 ns
  16777216 :   22.2 ns          /    25.9 ns
  33554432 :   56.5 ns          /    74.6 ns
  67108864 :   78.0 ns          /    96.2 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 :    1.0 ns          /     1.5 ns
    131072 :    1.5 ns          /     1.8 ns
    262144 :    1.8 ns          /     2.0 ns
    524288 :    2.1 ns          /     2.3 ns
   1048576 :    5.4 ns          /     7.2 ns
   2097152 :    7.1 ns          /     8.5 ns
   4194304 :    8.0 ns          /     8.9 ns
   8388608 :    9.3 ns          /     9.8 ns
  16777216 :   18.4 ns          /    21.1 ns
  33554432 :   53.2 ns          /    71.5 ns
  67108864 :   67.7 ns          /    83.5 ns

2×32GB DDR4 2666 MHz 16-18-18-35
FCLK @ 1800 MHz

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                                     :   9288.4 MB/s (1.0%)
 C copy backwards (32 byte blocks)                    :   9307.1 MB/s (1.1%)
 C copy backwards (64 byte blocks)                    :   9268.2 MB/s (1.1%)
 C copy                                               :   9321.3 MB/s (0.8%)
 C copy prefetched (32 bytes step)                    :   9526.6 MB/s
 C copy prefetched (64 bytes step)                    :   9618.8 MB/s (0.9%)
 C 2-pass copy                                        :   7782.8 MB/s (0.7%)
 C 2-pass copy prefetched (32 bytes step)             :   8521.4 MB/s (1.2%)
 C 2-pass copy prefetched (64 bytes step)             :   8532.7 MB/s (0.6%)
 C fill                                               :  11472.1 MB/s (1.1%)
 C fill (shuffle within 16 byte blocks)               :  11389.6 MB/s (1.3%)
 C fill (shuffle within 32 byte blocks)               :  11460.0 MB/s (1.7%)
 C fill (shuffle within 64 byte blocks)               :  11442.2 MB/s (0.7%)
 ---
 standard memcpy                                      :   8613.6 MB/s (1.7%)
 standard memset                                      :  13632.6 MB/s (1.8%)
 ---
 MOVSB copy                                           :   7508.9 MB/s (0.9%)
 MOVSD copy                                           :   7532.6 MB/s (1.3%)
 SSE2 copy                                            :   9977.8 MB/s (1.1%)
 SSE2 nontemporal copy                                :  16791.1 MB/s (1.4%)
 SSE2 copy prefetched (32 bytes step)                 :   9725.2 MB/s (1.1%)
 SSE2 copy prefetched (64 bytes step)                 :   9689.9 MB/s (1.0%)
 SSE2 nontemporal copy prefetched (32 bytes step)     :  17228.2 MB/s (1.2%)
 SSE2 nontemporal copy prefetched (64 bytes step)     :  17179.6 MB/s (1.0%)
 SSE2 2-pass copy                                     :   9214.5 MB/s (1.0%)
 SSE2 2-pass copy prefetched (32 bytes step)          :   9110.0 MB/s (1.1%)
 SSE2 2-pass copy prefetched (64 bytes step)          :   9147.5 MB/s (1.0%)
 SSE2 2-pass nontemporal copy                         :   4532.6 MB/s (0.6%)
 SSE2 fill                                            :  13015.2 MB/s (0.8%)
 SSE2 nontemporal fill                                :  28383.5 MB/s (0.5%)

==========================================================================
== 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 :    1.0 ns          /     1.4 ns
    131072 :    1.5 ns          /     1.8 ns
    262144 :    1.8 ns          /     2.0 ns
    524288 :    3.1 ns          /     3.6 ns
   1048576 :    6.7 ns          /     8.5 ns
   2097152 :    8.7 ns          /    10.1 ns
   4194304 :    9.9 ns          /    10.6 ns
   8388608 :   10.9 ns          /    11.1 ns
  16777216 :   23.8 ns          /    29.0 ns
  33554432 :   67.2 ns          /    91.1 ns
  67108864 :   89.0 ns          /   109.8 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 :    1.0 ns          /     1.4 ns
    131072 :    1.5 ns          /     1.8 ns
    262144 :    1.8 ns          /     1.9 ns
    524288 :    2.2 ns          /     2.4 ns
   1048576 :    5.3 ns          /     7.1 ns
   2097152 :    7.0 ns          /     8.5 ns
   4194304 :    8.0 ns          /     9.0 ns
   8388608 :    8.7 ns          /     9.1 ns
  16777216 :   14.0 ns          /    20.0 ns
  33554432 :   55.1 ns          /    77.9 ns
  67108864 :   74.4 ns          /    94.0 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 
Clone this wiki locally