Cold_Miserable

Worst case scenario for random access is a multiple level TLB miss, a memory refresh cycle and then a system management mode interrupt all occurring consecutively.

From the "analysis" I gather it works by encrypting the .exe and the key's are server-side. The hardware info is used to further encrypt it.

I think the goal should be to fool the checks rather than remove the encryption which would be a nightmare. CPUID can output whatever you want, it just reads MSR's. I'm sure there are possibilities to use kernel drivers to make windows functions also read out whatever you want.

The UK is a poor country with a money laundering capital tacked-on.

mov eax,64 lzcnt r8,rcx sub eax,r8d imul eax,1233 shr eax,12

Accurate to within 1.

Alderlake supports AVX512-FP16. Still only 9.6x faster than div. Most likely reciprocal is just too slow.

This is ~9.6x faster than "scalar".

ASM_TestDiv proc ;rcx out, rdx A, r8 B mov rax,05555555555555555H kmovq k1,rax vmovdqu8 zmm0,zmmword ptr [rdx] vmovdqu8 zmm4,zmmword ptr [r8] vpbroadcastw zmm3,word ptr [FLOAT16_F8] vmovdqu8 zmm2{k1},zmm0 ;lower 8-bit vmovdqu8 zmm16{k1},zmm4 ;lower 8-bit vpsrlw zmm1,zmm0,8 ;higher 8-bit vpsrlw zmm5,zmm4,8 ;higher 8-bit vpord zmm1,zmm1,zmm3 vpord zmm2,zmm2,zmm3 vpord zmm5,zmm5,zmm3 vpord zmm16,zmm16,zmm3 vsubph zmm1,zmm1,zmm3{rd-sae} ;fast conv 16FP vsubph zmm2,zmm2,zmm3{rd-sae} vsubph zmm5,zmm5,zmm3{ru-sae} vsubph zmm16,zmm16,zmm3{ru-sae} vrcpph zmm5,zmm5 vrcpph zmm16,zmm16 vfmadd213ph zmm1,zmm5,zmm3{rd-sae} vfmadd213ph zmm2,zmm16,zmm3{rd-sae} vxorpd zmm1,zmm1,zmm3 vxorpd zmm2,zmm2,zmm3 vpsllw zmm1,zmm1,8 vpord zmm1,zmm1,zmm2 vmovdqu8 zmmword ptr [rcx],zmm1 ;16 8-bit unsigned int ret

Heh? Surely fast convert 8-bit int to 16-bit FP,rcp+mul/div is a no-brainer? edit make that fast convert,rcp,fma (float 16 constant 1.0) and xor (same constant)

How is the built-in Administrator account faster than a named account? I wonder if its a bloatware effect. They need to test with all bloatware removed. No control flow guard, defender, firewall, appearance to performance, no sounds, no background, microcode dll's deleted, mitigations disabled in regedit, storport disabled, every service disabled, every app deleted, edge deleted etc.

Indeed. Intel is trash right now. No AVX512 thus AMD is infinitely faster.

Does Zen 5 have cldemote or senduipi ?

250 is nonsense. 2xFMA per cycle @ ~4.5Ghz = 32*4.5 = ~144 Gflops

Beating cuBlas is unlikely. You probably made a mistake. Last I tested it, it was even better than MKL in efficiency.

Its not possible to automate. Log is one example. You can't just use log(x) but log(x+1). There are other problems like the "zero problem" and its still possible to devise better approximations with other elementary operations such as |x| or ABS(x).

Intel makes money selling motherboard chipsets. There's no excuse, just greed.

Even a thicko can figure out this "research". You can't exercise your eyes to change their physical shape.

Biased article. 14nm was delayed and 10nm was delayed by 4 years.

??? The best action sci-fi ever made is a "goofy 80's movie".

TEC's are no good. The best bet is to bury copper tubing deep underground and have 2 loops.

Not interesting. Enter/leave also does the same thing as your save/restore rbp.

Far more interesting I recall there might be an instruction where rbp isn't allowed.

"experts in performance".

Windows uses unicode because NTFS uses unicode instead of efficient ascii.

Austrian organic dark chocolate granola is delectable.

Turning off "fullscreen optimizations" doesn't always work. It may not even be possible for DirectX 12. If you press Start and the taskbar appears, its not fullscreen. I wrote a simple Direct3D test program and entering fullscreen using alt-enter prevents start from appearing so DirectX 11.0 at least still supports fullscreen, for now.

I knew about the instruction that Daniel Lemire missed too. It seems useless to me.

US justice system failing again nicely I see. The average man would be rotting in jail and slowly wasting away.

Using fibre-optic cable where light travels in a zig-zag instead of using lasers increases latency too.

AVX512 is scarcely faster than radix at 1 million elements while being vastly more complicated.

They haven't even found an optimal sort for a ymm of 8 or zmm of 16 integers? Just 5 elements? That's completely and utterly useless.

Archimedean Dynasty and the Entrox drive is surely your first thought? That awesome underwater game made by BlueByte software.

No. AVX512 was disabled because the money-men want to force you to buy the far more expensive Sapphire-Rapids which are the same chip. Same old Intel. Learned nothing. They didn't just fuse it off, they also forced microcode updates that permanently disable it for working alderlakes.