Which is faster Win2k or Win98 For RC5 Cracking?

[WishMaster]

Hi, just a quick question on which of those are faster for cracking RC5. And Which OS is the fastest?

 

[Ken g6]

Klinux > Linux > Win2k > Win98se > Win98 (I think)

Also, I find it helps to set RC5's priority to 6, instead of 0, under Win98.

 

[jinsonxu]

How about WinMe? For a comparision, could someone else with a 900Mhz Duron using Win98, Win2K report the cracking speed? I'm getting 3.16MKeys/s in WinME. (Firewall on)

 

[jinsonxu]

By the way, could someone give a detailed explanation why the Athlon core performs better in RC5? I thought RC5 doesn't use the FPU? What kind of instructions are issued in cracking RC5?

Cause i thought that clock for clock, the PIII beat the Althon, unless the FPU was used?

There should be an explanation. For eg. It's my understanding that the G4 performs well in RC5 because it has more internal registers than the x86 class and thus the RC5 code can fit, eliminating the need for register renaming and the sorts. (This's my fuzzy grasp of the theory)

 

[sciencewhiz]

I read an interview of the person who wrote one of the earlier athlon cores and he said that the Athlon could execute more instructions per clock, if the program was optimized correctly.

RC5 depends heavily on rotates, so if you have a processor that is fast at that, it will be fast at RC5.

My experience is that all the operating systems benchmark at about the same rate. It is only in normal use where the graph the ken made is true.

 

[TuffGuy]

ddr

 

[eNeRGy]

<< Klinux > Linux > Win2k > Win98se > Win98 >>



About KLinux , I don;t know, but W2k is faster then Linux

Klinux? > Win2k > Linux > DOS > Win9x

 

[soni]

I have a 900MHz duron

It does 3.20Mkeys/sec
It uses 100MHz SDRAM Not 133MHz

I use Win2K Pro, 128MB, Duron 650@900

 

[jinsonxu]

Damn, it's true. Win 2K gives better results than WinME.

 

[sciencewhiz]

ram, fsb, amount of cache etc don't have anything to do with RC5 performance, it only depends on the integer portion of your CPU.

As for linux and 2k, It probably all depends on what you are running in the background.

 

[vss1980]

Actually, in all fairness NT4 is faster than Linux also at Rc5. I think that it is also around the same speed as Win2k (which is good considering win2k has more optimisations, etc.). As for Klinux, I dont know, but I doubt that it can offer more speed over standard Linux if it is using the same standard Linux kernal. The only advantage that it might have is that there are less processes running which would boost speed.

 

[bphantom]

It might be because the Win32 client is better optimized during the compiling process (someone mentioned something like this a while back), compared to the Linux client. My TBird does ~3.6 in W2K and ~3.5 in Linux (both 463 clients).

Brad..

 

[WishMaster]

thanks for the reply's

 

[SSP]

Cause i thought that clock for clock, the PIII beat the Althon, unless the FPU was used?

The Athlon will be on par or beat the P3 if there are no SSE optimizations (eg. Quake 3) in applications that require a fast FPU or interger calculations. When it comes to applications that are highly dependent on cache, the P3 (cumine) might do better (eg. Seti?). But the overall performance is very close to each other.

 

[vss1980]

Actually, the current T-bird Athlon is still at a disadvantage to the PIII Coppermine chips. The T-bird still only has 64-bit L2 cache interface. Even if it runs at the full processor speed this is still a disadvantage to the CuMine 256-bit cache.

This is where the AMD Mustang comes into play, making up this flaw of the current Athlon. At the moment clock-for-clock, they are very close, but the Mustang should change this.