Benchmarking a Windows 10 Home Laptop

I recently had the opportunity to benchmark two very similar Acer computers, differing mostly in the kind of Intel processor they shared. I found that only Novabench is worth using.

NovabenchInUseScore

As you may recall from my article on buying this so-called new laptop, I have an Acer laptop that’s cheap, heavy and running an i3 processor with 4GB RAM and (mostly) Windows 10 Home. I have replaced the slow HDD with a fast SSD Samsung 850 EVO drive.

As I helped a friend with some computer issues, I noticed that she had a very similar (but lighter) laptop (also Acer with 4GB RAM and same OS) with a less powerful N3060 processor. I was curious to see how the two computers performed in comparison, so I asked her to run some benchmarking software while using it and I did the same.

The quickest way to measure the performance of a Windows laptop is probably the “Windows Experience Index” which used to be made available in the “system” window, which one can bring up in several ways, such as pressing the Windows key + Pause/Break.

That is no longer shown in Windows 10, but it can be found by typing “winsat formal -v -xml c:\winstatresults.xml” in an admin command prompt. In my case, it ran for less than 2 minutes with a 4.4 system score.
  • Memory 5.9
  • CPU 6.8
  • CPU SubAgg 6.7
  • Video Encode 6.9
  • Graphics 4.4
  • Dx9, Dx10 9.9
  • Gaming 9.9
  • DiskScore 8.1
  • BildLab 14393.rs1_release.180209-1727

To my knowledge, the free benchmarking suites for Windows computers include Novabench, Prime 95, PCMark, 3DMark, SiSoft Sandra. I skipped PCMark and 3DMark as, at 2GB and, respectively, 4GB, they are far too hefty for what I need and decided to go with Prime95 as that seemed to be the only one that is also portable.

p95

I chose this one for the comparison as it seemed like a small download (from mersenne.org) and it was portable. However, the software belongs to a “prime search” project and their aim is to use your CPU cycles in a distributed model to search for such numbers. There is no bitcoin mining malware as far as I can tell – you have to press “Just Stress Testing” in the first window, then cancel the Torture Test offer and finally choose Benchmark under the Options menu.

Her results seem to indicate her laptoptablet being 4-5 times slower than mine. Her computer also had a McAfee antivirus as well as other Acer bloatware. I started to uninstall McAfee as the benchmark was running.
[Thu Feb 15 19:22:28 2018]
Compare your results to other computers at http://www.mersenne.org/report_benchmarks
Intel(R) Celeron(R) CPU  N3060  @ 1.60GHz
CPU speed: 2455.33 MHz, 2 cores
CPU features: Prefetchw, SSE, SSE2, SSE4
L1 cache size: 24 KB
L2 cache size: 1 MB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 256
Machine topology as determined by hwloc library:
 Machine#0 (total=1699980KB, Backend=Windows, hwlocVersion=1.11.9, ProcessName=prime95.exe)
  NUMANode#0 (local=1699980KB, total=1699980KB)
    Package#0 (CPUVendor=GenuineIntel, CPUFamilyNumber=6, CPUModelNumber=76, CPUModel="Intel(R) Celeron(R) CPU  N3060  @ 1.60GHz", CPUStepping=4)
      L2 (size=1024KB, linesize=64, ways=16, Inclusive=0)
        L1d (size=24KB, linesize=64, ways=6, Inclusive=0)
          Core (cpuset: 0x00000001)
            PU#0 (cpuset: 0x00000001)
      L2 (size=1024KB, linesize=64, ways=16, Inclusive=0)
        L1d (size=24KB, linesize=64, ways=6, Inclusive=0)
          Core (cpuset: 0x00000002)
            PU#1 (cpuset: 0x00000002)
Prime95 64-bit version 29.4, RdtscTiming=1
[Thu Feb 15 19:29:16 2018]
Timings for 2048K FFT length (2 cores, 1 worker): 9656.53 ms.  Throughput:  0.10 iter/sec.
Timings for 2048K FFT length (2 cores, 2 workers): 9529.85, 14250.22 ms.  Throughput:  0.18 iter/sec.
[Thu Feb 15 19:38:56 2018]
Timings for 2240K FFT length (2 cores, 1 worker): 130.66 ms.  Throughput:  7.65 iter/sec.
Timings for 2240K FFT length (2 cores, 2 workers): 412.91, 431.53 ms.  Throughput:  4.74 iter/sec.
Timings for 2304K FFT length (2 cores, 1 worker): 103.73 ms.  Throughput:  9.64 iter/sec.
Timings for 2304K FFT length (2 cores, 2 workers): 338.10, 188.68 ms.  Throughput:  8.26 iter/sec.
[Thu Feb 15 19:45:24 2018]
Timings for 2400K FFT length (2 cores, 1 worker): 358.48 ms.  Throughput:  2.79 iter/sec.
Timings for 2400K FFT length (2 cores, 2 workers): 151.39, 151.27 ms.  Throughput: 13.22 iter/sec.
Timings for 2560K FFT length (2 cores, 1 worker): 84.09 ms.  Throughput: 11.89 iter/sec.
Timings for 2560K FFT length (2 cores, 2 workers): 447.83, 454.03 ms.  Throughput:  4.44 iter/sec.
Timings for 2688K FFT length (2 cores, 1 worker): 87.25 ms.  Throughput: 11.46 iter/sec.
Timings for 2688K FFT length (2 cores, 2 workers): 508.98, 439.00 ms.  Throughput:  4.24 iter/sec.
Timings for 2880K FFT length (2 cores, 1 worker): 118.63 ms.  Throughput:  8.43 iter/sec.
Timings for 2880K FFT length (2 cores, 2 workers): 168.91, 170.23 ms.  Throughput: 11.79 iter/sec.
Timings for 3072K FFT length (2 cores, 1 worker): 362.37 ms.  Throughput:  2.76 iter/sec.
Timings for 3072K FFT length (2 cores, 2 workers): 235.20, 224.75 ms.  Throughput:  8.70 iter/sec.
Timings for 3200K FFT length (2 cores, 1 worker): 168.99 ms.  Throughput:  5.92 iter/sec.
Timings for 3200K FFT length (2 cores, 2 workers): 175.26, 176.23 ms.  Throughput: 11.38 iter/sec.
Timings for 3360K FFT length (2 cores, 1 worker): 179.91 ms.  Throughput:  5.56 iter/sec.
Timings for 3360K FFT length (2 cores, 2 workers): 271.94, 271.40 ms.  Throughput:  7.36 iter/sec.
[Thu Feb 15 19:50:32 2018]
Timings for 3584K FFT length (2 cores, 1 worker): 100.17 ms.  Throughput:  9.98 iter/sec.
Timings for 3584K FFT length (2 cores, 2 workers): 194.70, 193.35 ms.  Throughput: 10.31 iter/sec.
Timings for 3840K FFT length (2 cores, 1 worker): 122.18 ms.  Throughput:  8.18 iter/sec.
Timings for 3840K FFT length (2 cores, 2 workers): 257.53, 247.49 ms.  Throughput:  7.92 iter/sec.
Timings for 4000K FFT length (2 cores, 1 worker): 147.85 ms.  Throughput:  6.76 iter/sec.
Timings for 4000K FFT length (2 cores, 2 workers): 232.22, 227.61 ms.  Throughput:  8.70 iter/sec.
Timings for 4096K FFT length (2 cores, 1 worker): 116.24 ms.  Throughput:  8.60 iter/sec.
Timings for 4096K FFT length (2 cores, 2 workers): 233.60, 233.15 ms.  Throughput:  8.57 iter/sec.
Timings for 4480K FFT length (2 cores, 1 worker): 156.82 ms.  Throughput:  6.38 iter/sec.
Timings for 4480K FFT length (2 cores, 2 workers): 370.75, 343.89 ms.  Throughput:  5.61 iter/sec.
Timings for 4608K FFT length (2 cores, 1 worker): 144.07 ms.  Throughput:  6.94 iter/sec.
Timings for 4608K FFT length (2 cores, 2 workers): 325.57, 313.50 ms.  Throughput:  6.26 iter/sec.
Timings for 4800K FFT length (2 cores, 1 worker): 156.85 ms.  Throughput:  6.38 iter/sec.
Timings for 4800K FFT length (2 cores, 2 workers): 282.22, 280.27 ms.  Throughput:  7.11 iter/sec.
Timings for 5120K FFT length (2 cores, 1 worker): 134.45 ms.  Throughput:  7.44 iter/sec.
Timings for 5120K FFT length (2 cores, 2 workers): 339.50, 327.89 ms.  Throughput:  6.00 iter/sec.
[Thu Feb 15 19:55:51 2018]
Timings for 5376K FFT length (2 cores, 1 worker): 156.06 ms.  Throughput:  6.41 iter/sec.
Timings for 5376K FFT length (2 cores, 2 workers): 364.78, 358.82 ms.  Throughput:  5.53 iter/sec.
Timings for 5600K FFT length (2 cores, 1 worker): 185.44 ms.  Throughput:  5.39 iter/sec.
Timings for 5600K FFT length (2 cores, 2 workers): 355.94, 349.42 ms.  Throughput:  5.67 iter/sec.
Timings for 5760K FFT length (2 cores, 1 worker): 188.21 ms.  Throughput:  5.31 iter/sec.
Timings for 5760K FFT length (2 cores, 2 workers): 378.64, 368.21 ms.  Throughput:  5.36 iter/sec.
Timings for 6144K FFT length (2 cores, 1 worker): 181.61 ms.  Throughput:  5.51 iter/sec.
Timings for 6144K FFT length (2 cores, 2 workers): 448.09, 419.93 ms.  Throughput:  4.61 iter/sec.
Timings for 6400K FFT length (2 cores, 1 worker): 221.75 ms.  Throughput:  4.51 iter/sec.
Timings for 6400K FFT length (2 cores, 2 workers): 387.27, 381.72 ms.  Throughput:  5.20 iter/sec.
Timings for 6720K FFT length (2 cores, 1 worker): 370.59 ms.  Throughput:  2.70 iter/sec.
Timings for 6720K FFT length (2 cores, 2 workers): 564.25, 541.92 ms.  Throughput:  3.62 iter/sec.
Timings for 6912K FFT length (2 cores, 1 worker): 252.77 ms.  Throughput:  3.96 iter/sec.
Timings for 6912K FFT length (2 cores, 2 workers): 571.20, 539.58 ms.  Throughput:  3.60 iter/sec.
[Thu Feb 15 20:01:07 2018]
Timings for 7168K FFT length (2 cores, 1 worker): 261.56 ms.  Throughput:  3.82 iter/sec.
Timings for 7168K FFT length (2 cores, 2 workers): 1039.39, 937.12 ms.  Throughput:  2.03 iter/sec.
Timings for 7680K FFT length (2 cores, 1 worker): 303.49 ms.  Throughput:  3.29 iter/sec.
Timings for 7680K FFT length (2 cores, 2 workers): 617.53, 598.68 ms.  Throughput:  3.29 iter/sec.
Timings for 8192K FFT length (2 cores, 1 worker): 336.67 ms.  Throughput:  2.97 iter/sec.
Timings for 8192K FFT length (2 cores, 2 workers): 565.53, 537.08 ms.  Throughput:  3.63 iter/sec.

It is a bit difficult interpreting these numbers, but from the list of results (mersenne-results), the one closest to mine would be i3-4005U. Anyway, the numbers (mine are below) do indeed match the subjective observations, as her computer seemed extremely slow.

In contrast, my own laptop had a significantly better processor (though obsolete by today’s standards), no McAfee and a more or less clean Windows 10 install. Mine is updated, with the “creators update” blocked. However, the computer was in use while the benchmark was running with a Chrome-based browser with about 10 extensions and with two windows having a total of about 20 tabs, VLC playing a SomaFM radio stream, and most software running off a Veracrypt-ed drive; also f.lux.

[Thu Feb 15 19:21:46 2018]
Compare your results to other computers at http://www.mersenne.org/report_benchmarks
Intel(R) Core(TM) i3-4030U CPU @ 1.90GHz
CPU speed: 1895.77 MHz, 2 hyperthreaded cores
CPU features: Prefetch, SSE, SSE2, SSE4, AVX, AVX2, FMA
L1 cache size: 32 KB
L2 cache size: 256 KB, L3 cache size: 3 MB
L1 cache line size: 64 bytes
L2 cache line size: 64 bytes
TLBS: 64
Machine topology as determined by hwloc library:
Machine#0 (total=1691468KB, Backend=Windows, hwlocVersion=1.11.9, ProcessName=prime95.exe)
  NUMANode#0 (local=1691468KB, total=1691468KB)
    Package#0 (CPUVendor=GenuineIntel, CPUFamilyNumber=6, CPUModelNumber=69, CPUModel="Intel(R) Core(TM) i3-4030U CPU @ 1.90GHz", CPUStepping=1)
      L3 (size=3072KB, linesize=64, ways=12, Inclusive=1)
        L2 (size=256KB, linesize=64, ways=8, Inclusive=0)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x00000003)
              PU#0 (cpuset: 0x00000001)
              PU#1 (cpuset: 0x00000002)
        L2 (size=256KB, linesize=64, ways=8, Inclusive=0)
          L1d (size=32KB, linesize=64, ways=8, Inclusive=0)
            Core (cpuset: 0x0000000c)
              PU#2 (cpuset: 0x00000004)
              PU#3 (cpuset: 0x00000008)
Prime95 64-bit version 29.4, RdtscTiming=1
Timings for 2048K FFT length (2 cores, 1 worker): 10.36 ms.  Throughput: 96.56 iter/sec.
Timings for 2048K FFT length (2 cores, 2 workers): 21.97, 21.32 ms.  Throughput: 92.41 iter/sec.
Timings for 2048K FFT length (2 cores hyperthreaded, 1 worker): 11.28 ms.  Throughput: 88.66 iter/sec.
Timings for 2048K FFT length (2 cores hyperthreaded, 2 workers): 21.38, 21.01 ms.  Throughput: 94.35 iter/sec.
Timings for 2304K FFT length (2 cores, 1 worker): 11.80 ms.  Throughput: 84.74 iter/sec.
Timings for 2304K FFT length (2 cores, 2 workers): 21.82, 21.68 ms.  Throughput: 91.96 iter/sec.
Timings for 2304K FFT length (2 cores hyperthreaded, 1 worker): 11.92 ms.  Throughput: 83.86 iter/sec.
Timings for 2304K FFT length (2 cores hyperthreaded, 2 workers): 29.62, 28.37 ms.  Throughput: 69.01 iter/sec.
Timings for 2400K FFT length (2 cores, 1 worker): 13.31 ms.  Throughput: 75.13 iter/sec.
Timings for 2400K FFT length (2 cores, 2 workers): 24.88, 24.38 ms.  Throughput: 81.22 iter/sec.
Timings for 2400K FFT length (2 cores hyperthreaded, 1 worker): 13.56 ms.  Throughput: 73.73 iter/sec.
Timings for 2400K FFT length (2 cores hyperthreaded, 2 workers): 25.60, 24.21 ms.  Throughput: 80.37 iter/sec.
Timings for 2560K FFT length (2 cores, 1 worker): 13.54 ms.  Throughput: 73.84 iter/sec.
Timings for 2560K FFT length (2 cores, 2 workers): 26.02, 25.80 ms.  Throughput: 77.19 iter/sec.
Timings for 2560K FFT length (2 cores hyperthreaded, 1 worker): 13.16 ms.  Throughput: 75.96 iter/sec.
Timings for 2560K FFT length (2 cores hyperthreaded, 2 workers): 30.54, 28.47 ms.  Throughput: 67.87 iter/sec.
Timings for 2688K FFT length (2 cores, 1 worker): 13.35 ms.  Throughput: 74.88 iter/sec.
Timings for 2688K FFT length (2 cores, 2 workers): 26.22, 25.88 ms.  Throughput: 76.79 iter/sec.
[Thu Feb 15 19:26:56 2018]
Timings for 2688K FFT length (2 cores hyperthreaded, 1 worker): 14.45 ms.  Throughput: 69.19 iter/sec.
Timings for 2688K FFT length (2 cores hyperthreaded, 2 workers): 34.31, 29.35 ms.  Throughput: 63.22 iter/sec.
Timings for 2880K FFT length (2 cores, 1 worker): 15.37 ms.  Throughput: 65.06 iter/sec.
Timings for 2880K FFT length (2 cores, 2 workers): 36.75, 35.41 ms.  Throughput: 55.45 iter/sec.
Timings for 2880K FFT length (2 cores hyperthreaded, 1 worker): 18.42 ms.  Throughput: 54.30 iter/sec.
Timings for 2880K FFT length (2 cores hyperthreaded, 2 workers): 37.90, 34.74 ms.  Throughput: 55.17 iter/sec.
Timings for 3072K FFT length (2 cores, 1 worker): 26.46 ms.  Throughput: 37.79 iter/sec.
Timings for 3072K FFT length (2 cores, 2 workers): 37.53, 36.41 ms.  Throughput: 54.11 iter/sec.
Timings for 3072K FFT length (2 cores hyperthreaded, 1 worker): 17.55 ms.  Throughput: 56.98 iter/sec.
Timings for 3072K FFT length (2 cores hyperthreaded, 2 workers): 34.64, 31.51 ms.  Throughput: 60.60 iter/sec.
Timings for 3200K FFT length (2 cores, 1 worker): 16.18 ms.  Throughput: 61.79 iter/sec.
Timings for 3200K FFT length (2 cores, 2 workers): 32.21, 31.97 ms.  Throughput: 62.33 iter/sec.
Timings for 3200K FFT length (2 cores hyperthreaded, 1 worker): 17.76 ms.  Throughput: 56.31 iter/sec.
Timings for 3200K FFT length (2 cores hyperthreaded, 2 workers): 38.39, 34.44 ms.  Throughput: 55.09 iter/sec.
Timings for 3360K FFT length (2 cores, 1 worker): 23.22 ms.  Throughput: 43.06 iter/sec.
Timings for 3360K FFT length (2 cores, 2 workers): 37.66, 36.75 ms.  Throughput: 53.76 iter/sec.
Timings for 3360K FFT length (2 cores hyperthreaded, 1 worker): 19.14 ms.  Throughput: 52.26 iter/sec.
Timings for 3360K FFT length (2 cores hyperthreaded, 2 workers): 41.54, 40.29 ms.  Throughput: 48.89 iter/sec.
[Thu Feb 15 19:32:01 2018]
Timings for 3456K FFT length (2 cores, 1 worker): 19.38 ms.  Throughput: 51.61 iter/sec.
Timings for 3456K FFT length (2 cores, 2 workers): 48.18, 46.05 ms.  Throughput: 42.47 iter/sec.
Timings for 3456K FFT length (2 cores hyperthreaded, 1 worker): 20.68 ms.  Throughput: 48.36 iter/sec.
Timings for 3456K FFT length (2 cores hyperthreaded, 2 workers): 45.81, 41.48 ms.  Throughput: 45.93 iter/sec.
Timings for 3584K FFT length (2 cores, 1 worker): 19.09 ms.  Throughput: 52.38 iter/sec.
Timings for 3584K FFT length (2 cores, 2 workers): 38.94, 41.60 ms.  Throughput: 49.72 iter/sec.
Timings for 3584K FFT length (2 cores hyperthreaded, 1 worker): 19.18 ms.  Throughput: 52.14 iter/sec.
Timings for 3584K FFT length (2 cores hyperthreaded, 2 workers): 40.26, 38.50 ms.  Throughput: 50.82 iter/sec.
Timings for 3840K FFT length (2 cores, 1 worker): 20.35 ms.  Throughput: 49.15 iter/sec.
Timings for 3840K FFT length (2 cores, 2 workers): 39.43, 39.08 ms.  Throughput: 50.95 iter/sec.
Timings for 3840K FFT length (2 cores hyperthreaded, 1 worker): 20.47 ms.  Throughput: 48.86 iter/sec.
Timings for 3840K FFT length (2 cores hyperthreaded, 2 workers): 41.40, 40.15 ms.  Throughput: 49.06 iter/sec.
Timings for 4096K FFT length (2 cores, 1 worker): 21.71 ms.  Throughput: 46.06 iter/sec.
Timings for 4096K FFT length (2 cores, 2 workers): 44.02, 43.09 ms.  Throughput: 45.92 iter/sec.
Timings for 4096K FFT length (2 cores hyperthreaded, 1 worker): 23.44 ms.  Throughput: 42.65 iter/sec.
Timings for 4096K FFT length (2 cores hyperthreaded, 2 workers): 46.95, 47.38 ms.  Throughput: 42.41 iter/sec.
Timings for 4480K FFT length (2 cores, 1 worker): 22.94 ms.  Throughput: 43.60 iter/sec.
Timings for 4480K FFT length (2 cores, 2 workers): 45.79, 45.05 ms.  Throughput: 44.03 iter/sec.
[Thu Feb 15 19:37:06 2018]
Timings for 4480K FFT length (2 cores hyperthreaded, 1 worker): 25.44 ms.  Throughput: 39.30 iter/sec.
Timings for 4480K FFT length (2 cores hyperthreaded, 2 workers): 59.27, 52.06 ms.  Throughput: 36.08 iter/sec.
Timings for 4608K FFT length (2 cores, 1 worker): 24.62 ms.  Throughput: 40.62 iter/sec.
Timings for 4608K FFT length (2 cores, 2 workers): 47.20, 46.67 ms.  Throughput: 42.62 iter/sec.
Timings for 4608K FFT length (2 cores hyperthreaded, 1 worker): 24.44 ms.  Throughput: 40.92 iter/sec.
Timings for 4608K FFT length (2 cores hyperthreaded, 2 workers): 50.59, 48.48 ms.  Throughput: 40.40 iter/sec.
Timings for 4800K FFT length (2 cores, 1 worker): 24.46 ms.  Throughput: 40.89 iter/sec.
Timings for 4800K FFT length (2 cores, 2 workers): 49.00, 48.51 ms.  Throughput: 41.02 iter/sec.
Timings for 4800K FFT length (2 cores hyperthreaded, 1 worker): 26.16 ms.  Throughput: 38.23 iter/sec.
Timings for 4800K FFT length (2 cores hyperthreaded, 2 workers): 54.32, 52.00 ms.  Throughput: 37.64 iter/sec.
Timings for 5120K FFT length (2 cores, 1 worker): 27.35 ms.  Throughput: 36.57 iter/sec.
Timings for 5120K FFT length (2 cores, 2 workers): 55.45, 54.00 ms.  Throughput: 36.55 iter/sec.
Timings for 5120K FFT length (2 cores hyperthreaded, 1 worker): 34.04 ms.  Throughput: 29.38 iter/sec.
Timings for 5120K FFT length (2 cores hyperthreaded, 2 workers): 71.92, 65.60 ms.  Throughput: 29.15 iter/sec.
Timings for 5376K FFT length (2 cores, 1 worker): 27.81 ms.  Throughput: 35.96 iter/sec.
Timings for 5376K FFT length (2 cores, 2 workers): 55.39, 54.55 ms.  Throughput: 36.39 iter/sec.
Timings for 5376K FFT length (2 cores hyperthreaded, 1 worker): 51.42 ms.  Throughput: 19.45 iter/sec.
Timings for 5376K FFT length (2 cores hyperthreaded, 2 workers): 74.10, 64.40 ms.  Throughput: 29.02 iter/sec.
[Thu Feb 15 19:42:23 2018]
Timings for 5760K FFT length (2 cores, 1 worker): 33.79 ms.  Throughput: 29.59 iter/sec.
Timings for 5760K FFT length (2 cores, 2 workers): 62.32, 62.08 ms.  Throughput: 32.16 iter/sec.
Timings for 5760K FFT length (2 cores hyperthreaded, 1 worker): 35.52 ms.  Throughput: 28.16 iter/sec.
Timings for 5760K FFT length (2 cores hyperthreaded, 2 workers): 73.59, 71.23 ms.  Throughput: 27.63 iter/sec.
Timings for 6144K FFT length (2 cores, 1 worker): 32.51 ms.  Throughput: 30.76 iter/sec.
Timings for 6144K FFT length (2 cores, 2 workers): 66.57, 65.24 ms.  Throughput: 30.35 iter/sec.
Timings for 6144K FFT length (2 cores hyperthreaded, 1 worker): 37.36 ms.  Throughput: 26.76 iter/sec.
Timings for 6144K FFT length (2 cores hyperthreaded, 2 workers): 93.22, 82.81 ms.  Throughput: 22.80 iter/sec.
Timings for 6400K FFT length (2 cores, 1 worker): 37.95 ms.  Throughput: 26.35 iter/sec.
Timings for 6400K FFT length (2 cores, 2 workers): 72.31, 70.17 ms.  Throughput: 28.08 iter/sec.
Timings for 6400K FFT length (2 cores hyperthreaded, 1 worker): 36.02 ms.  Throughput: 27.77 iter/sec.
Timings for 6400K FFT length (2 cores hyperthreaded, 2 workers): 73.35, 71.12 ms.  Throughput: 27.69 iter/sec.
Timings for 6720K FFT length (2 cores, 1 worker): 39.23 ms.  Throughput: 25.49 iter/sec.
Timings for 6720K FFT length (2 cores, 2 workers): 77.51, 75.54 ms.  Throughput: 26.14 iter/sec.
Timings for 6720K FFT length (2 cores hyperthreaded, 1 worker): 49.41 ms.  Throughput: 20.24 iter/sec.
Timings for 6720K FFT length (2 cores hyperthreaded, 2 workers): 105.57, 101.03 ms.  Throughput: 19.37 iter/sec.
Timings for 6912K FFT length (2 cores, 1 worker): 40.68 ms.  Throughput: 24.58 iter/sec.
[Thu Feb 15 19:47:33 2018]
Timings for 6912K FFT length (2 cores, 2 workers): 76.63, 74.99 ms.  Throughput: 26.38 iter/sec.
Timings for 6912K FFT length (2 cores hyperthreaded, 1 worker): 43.93 ms.  Throughput: 22.76 iter/sec.
Timings for 6912K FFT length (2 cores hyperthreaded, 2 workers): 88.79, 87.56 ms.  Throughput: 22.68 iter/sec.
Timings for 7168K FFT length (2 cores, 1 worker): 41.13 ms.  Throughput: 24.32 iter/sec.
Timings for 7168K FFT length (2 cores, 2 workers): 82.81, 82.11 ms.  Throughput: 24.25 iter/sec.
Timings for 7168K FFT length (2 cores hyperthreaded, 1 worker): 52.75 ms.  Throughput: 18.96 iter/sec.
Timings for 7168K FFT length (2 cores hyperthreaded, 2 workers): 111.52, 109.18 ms.  Throughput: 18.13 iter/sec.
Timings for 7680K FFT length (2 cores, 1 worker): 41.48 ms.  Throughput: 24.11 iter/sec.
Timings for 7680K FFT length (2 cores, 2 workers): 85.59, 81.63 ms.  Throughput: 23.94 iter/sec.
Timings for 7680K FFT length (2 cores hyperthreaded, 1 worker): 49.10 ms.  Throughput: 20.37 iter/sec.
Timings for 7680K FFT length (2 cores hyperthreaded, 2 workers): 100.16, 97.15 ms.  Throughput: 20.28 iter/sec.
Timings for 8064K FFT length (2 cores, 1 worker): 45.93 ms.  Throughput: 21.77 iter/sec.
Timings for 8064K FFT length (2 cores, 2 workers): 90.65, 89.26 ms.  Throughput: 22.23 iter/sec.
Timings for 8064K FFT length (2 cores hyperthreaded, 1 worker): 60.77 ms.  Throughput: 16.46 iter/sec.
Timings for 8064K FFT length (2 cores hyperthreaded, 2 workers): 127.90, 124.89 ms.  Throughput: 15.83 iter/sec.
Timings for 8192K FFT length (2 cores, 1 worker): 50.94 ms.  Throughput: 19.63 iter/sec.
Timings for 8192K FFT length (2 cores, 2 workers): 102.44, 101.67 ms.  Throughput: 19.60 iter/sec.
[Thu Feb 15 19:52:50 2018]
Timings for 8192K FFT length (2 cores hyperthreaded, 1 worker): 60.99 ms.  Throughput: 16.40 iter/sec.
Timings for 8192K FFT length (2 cores hyperthreaded, 2 workers): 136.21, 127.74 ms.  Throughput: 15.17 iter/sec.
[Fri Feb 16 17:39:48 2018]
Quitting GIMPS after current work completes.

A comparison between the i3 and Pentium (e.g., gadgets-compare) is rather subjective, as often the exact version compared might not match what one has, as it happens to these newer ones.

On Quora, somebody compares the newer i3-7100 and the G4620 (quora-comp).

At this moment (December 2017) I’m not sure there really is much difference. The Pentium G4620 is based on the Kaby Lake architecture, just like the 7-series i3s. Just like the 7-series i3s the Pentium has Hyper Threading and two processing cores for four threads. They both have intel 630 integrated graphics and both use identical 1151 sockets. In fact the only real difference between the two is that the i3 supports AVX2.0 while the Pentium has no AVX support and the Pentium has a lower core clock.

As a point of interest UserBenchmark places the performance of the i3–7100 as about 4–5% greater than the Pentium G4620:

Intel Core i3-7100 vs Pentium G4620

Interestingly, the i3–7100 clock is 3.9ghz, which is about 5% faster than the G4620’s 3.7ghz core clock. So the Pentium’s performance is directly inline with what you’d expect of the same generation i3 with the same clock rate.

It’s my belief that the current crop of Pentium processors are simply i3s that failed to test well enough after manufacture for the i3 brand. Thus they’re sold for a lower price with some features turned off and a lower clock rate. This is a fairly standard practice in CPU marketing and development.

CrystalMarkSSD850EVOOne significant improvement I have made to my cheap Acer i3 laptop is the 850 EVO SSD which compares very favourably to the previous ST500LT012.

novabench

Novabench is a small download, it completes in 1h or less and can be uploaded to their website.

novabench666I re-ran Novabench, (result above, first as shown in the first screenshot,) without the browser running, but still with Veracrypt and other things on, getting a substantial score improvement. I am keeping this one as it fits my requirements.

7zip

benchmrk7zip, an archiver, has a benchmarking subroutine (under the Tools menu), though apart from the forum linked under the image above, a list of results is hard to come by.

The help section explains the numbers and to me it doesn’t seem very rigorous if you want to use it for comparison benchmarking.

There are two tests:

  1. Compression with LZMA method
  2. Decompression with LZMA method

The benchmark shows a rating in MIPS (million instructions per second). The rating value is calculated from the measured speed, and it is normalized with results of Intel Core 2 CPU with multi-threading option switched off. So if you have modern CPU from Intel or AMD, rating values in single-thread mode must be close to real CPU frequency.

You can change the dictionary size to increase memory usage. Also you can change the number of threads.

The CPU Usage column shows the percentage of time the processor is working. It's normalized for a one-thread load. For example, 180% CPU Usage for 2 threads can mean that average CPU usage is about 90% for each thread.

The Rating / Usage column shows rating normalized for 100% of CPU usage. That column shows performance of the one CPU thread. It must be close to real CPU frequency, if you have modern CPU.

The Total rating shows averages of the compressing and decompression ratings.

Compression speed and rating strongly depend from memory (RAM) latency.

Decompression speed and rating strongly depend on CPU integer operations. For example, an Intel Pentium 4 has big branch misprediction penalty (which is effect of long pipeline) and pretty slow multiply and shift operations. So, the Pentium 4 has pretty low decompressing ratings.

sandra

Sisoftware Sandra is a slightly more complex testing suite, using a client-server model, which makes it better suited for LANs.

I found the stand-alone install difficult to use and a bit buggy but it identified and reported on the hardware well.

SiSoftware Sandra - abbreviated results, repeated data deleted

Processor Multi-Media
Analysing...
Aggregated Score : 79.65MPix/s
Result ID : Intel(R) Core(TM) i3-4030U CPU @ 1.90GHz (2C 4T 1.9GHz, 1.9GHz IMC, 2x 256kB L2, 3MB L3)
Speed : 1896MHz
Capacity : 4Unit(s)
Power : 25.00W
Finished Successfully : Yes

Processor Cryptography : 2.232GB/s

Processor Financial Analysis : 5.13kOPT/s

Processor Scientific Analysis : 6.74GFLOPS

.NET Arithmetic : 8.46GOPS

Memory Bandwidth : 7.399GB/s

Cache & Memory Latency : 65.4ns

File System Bandwidth
Error (5202) : EvoAcr (NTFS, 4kB) : Read/Write Performance : 11374 MB free space required on device \\.\C:\!
Aggregated Score : 177.758MB/s
Result ID : Samsung SSD 850 EVO 500GB (500.1GB, SATA600, 2.5", SSD, OPAL) (NTFS, 4kB)
Capacity : 500.11GB
Finished Successfully : Yes

File System I/O : 8652.7IOPS

GP (GPU/CPU/APU) Processing : 76.71MPix/s
Result ID : Intel(R) HD Graphics Family (160SP 20C 1GHz, 256kB L2, 1.6GB DDR3 1.6GHz 64-bit, Integrated Graphics) (OpenCL)
Speed : 1000MHz
Capacity : 20Unit(s)
Power : 25.00W
Finished Successfully : Yes

GP (GPU/CPU/APU) Cryptography : 0.845GB/s

GP (GPU/CPU/APU) Financial Analysis : 88.36kOPT/s

GP (GPU/CPU/APU) Scientific Analysis
Analysing...
Error (339) : OpenCL : Intel(R) HD Graphics Family (160SP 20C 1GHz, 256kB L2, 1.6GB DDR3 1.6GHz 64-bit) : Floating-Point (Normal/Single Precision) : GP(GPU) call failed. Try another interface (e.g. OpenCL/ComputeShader/CUDA/etc.) or update video drivers.
Error (339) : D3D 11 : Intel(R) HD Graphics Family (160SP 20C 1GHz, 512kB L2, 128MB DDR3 1.6GHz 64-bit) : Floating-Point (Normal/Single Precision) : GP(GPU) call failed. Try another interface (e.g. OpenCL/ComputeShader/CUDA/etc.) or update video drivers.
Error (339) : D3D 11 : Intel(R) HD Graphics Family (160SP 20C 1GHz, 512kB L2, 128MB DDR3 1.6GHz 64-bit) : Floating-Point (High/Double Precision) : GP(GPU) call failed. Try another interface (e.g. OpenCL/ComputeShader/CUDA/etc.) or update video drivers.
Finished Successfully : No

Processor Scientific Analysis : 6.59GFLOPS

Media (Audio/Video) Transcode : 5.950MB/s
Result ID : Intel(R) HD Graphics Family (20CU 160SP  1GHz, 512kB L2, 128MB DDR3 1.6GHz 64-bit, Integrated Graphics) (Microsoft H264 Video Decoder MFT > Intel® Quick Sync Video H.264 Encoder MFT; Microsoft AAC Audio Decoder MFT > Microsoft AAC Audio Encoder MFT)
Speed : 1000MHz
Capacity : 1Unit(s)
Power : 25.00W
Finished Successfully : Yes

GP (GPU/CPU/APU) Memory Bandwidth : 6.082GB/s
Result ID : Intel(R) HD Graphics Family (160SP 20C 1GHz, 256kB L2, 1.6GB DDR3 1.6GHz 64-bit, Integrated Graphics) (OpenCL)
Speed : 1600MHz
Capacity : 1600MB
Power : 25.00W
Finished Successfully : Yes

GP (GPU/CPU/APU) Memory Latency : 173.3ns

Overall Score
Aggregated Score : 3.70kPT
Results Interpretation : Higher Scores mean Better Performance.
Decimal Numeral System (base 10) : 1GPT = 1000MPT, 1MPT = 1000kPT, 1kPT = 1000PT, etc.
Result ID : Acer Aspire E5-471P Type1Family (Acer EA40_HB) (Intel Core i3-4030U CPU @ 1.90GHz; Intel Core DRAM Controller; 4GB Hynix HMT451S6BFR8A-PB DDR3 SO; Samsung SSD 850 EVO 500GB; Intel HD Graphics Family)
Finished Successfully : Yes

Overall Score1aSandra does provide extensive reporting and comparison with other similar hardware, but in my experience the similar hardware wasn’t exactly all that similar. I had to uncheck the pre-checked i7 models and check the only two i3 CPUs in the list.

rankerThough the website is still updated, the “ranker” site appears to be currently down.

I uninstalled it and it offered a license in return for completing some tasks – these involved depositing money with a betting site and buying fast food gift cards, so I passed.

Most buying advice no longer considers i3 processors, recommending an Atom or N#### class processor for budget conscious buyers, and i5-i7 for the rest (lptp-cpu).

In conclusion, I recommend Novabench as the only free benchmarking software worth trying. If, like me, you are using an encrypted drive, you might want to install it on the main (unencrypted) drive, so that you can test without the encryption software running.

Sources / More info: mersenne-results, quora-comp, gadgets-compare, wiki-compare, lptp-cpu

blog comments powered by Disqus
Your gift enables us to continue to provide high quality reporting with little or no advertising.