Several tests of the Sparks R500 and R1000 performance from the MCST
Luckily, I got into the hands of a console from two sparks of the MCST - R500 and R1000. I admit I'm rather skeptical about the projects of Russian electronic engineers, so I wanted to look at their performance. The results of which I hasten to share.As tests, I used runs of guest encryption algorithms - GOST R34.10-2001, GOST R34.11-94, GOST 28147-89 in one stream. Encryption and hash were run on 32k blocks. As you can see from the graphs, the performance of the experimental R1000 800Mhz corresponds to the ARM Cortex A9 1.2Ghz, which of course was pleasantly surprised. I have not yet been able to access the GHz version of R1000, there I hope the performance will be higher. Here is the value in megabits / seconds:
r500-500Mhz 38 30 28 14 8 6 6
r1000-800Mhz 80 70 68 38 22 16 18
PB A9 1200Mhz 86 78 74 36 38 30 32
x86 vm 226 216 210 122 194 150 164
That's all for now. Next time I'll see how the 4x4 matrix parallelizes and, if it works out, I’ll drive on Elbrus. Of course, I understand that the tests are extremely subjective and one-sided, but here I am in the heat of the heat.