Little more then 2 years left.

[murdmath]

I noticed that on the stats that at our current rate we have little more then 2 years to go. We all know that faster processors are coming out that could increase each of our keyrates, but we all know that the P4 sucks at RC5. Do you think that this will slow down the rate at which the time has been decreasing? I hope you can understand what i'm trying to get at.

Murdmath

 

[IJump]

The new prerelease client for RC5 is better for the P4. That will help. Plus with all of the new people joining (I assume this is probably happening on all teams), the rate of decrease of remaining time should just increase. Hopefully, this is at least as clear as your question.

 

[JimMc]

Of course it is possible that RC5 could also end any day .

Hard to predict the impact of the P4--yeah there will be more of them, but they may also be replacing K6 300's, Cyrix, etc. chips which suck even more.

 

[networkman]

It's also possible that the key was already found, but lost in a corrupted packet! :Q :Q

 

[phatstyl]

<< It's also possible that the key was already found, but lost in a corrupted packet! >>


DOH!

 

[xtreme2k]

I must say we should be finished in about 9 months. Considering the CPU power keeps improving and hoping we dont have to crack till the last block to get the key.

 

[Bleep]

Wouldn't it be a real kick in the butt if all the keys were exhausted and the code still not cracked because of a programming error? HOLY MOLIE what a terrible thought.

Bleep

 

[SlangNRox]

It will be nice when rc5 is finished so we can move onto more useful distributed projects. I know I'm not going to abandon rc5 because I've spent ove 2 years cracking and now that I'm cracking at a decent rate I'm finally moving up in the stats.

 

[AppleTalking]

RC5 is going to end?! Where will I go, what will I do? <sobs>



Nick

 

[murdmath]

<< I must say we should be finished in about 9 months. Considering the CPU power keeps improving and hoping we dont have to crack till the last block to get the key. >>



Ahh, That is my point. Will the P4 decrease our current rate of cpu power growth. Will the curve be less steep because each P4 is actually weaker per cpu then our current cpu's RC5 wise.

murdmath

 

[MikeyP]

I haven't kept up on it, but is it possible to bitslice using SSE2. If so, the P4 could be about as fast as the G4 clock/clock. That is scary!

 

[ViRGE]

Assuming that older machines are replaced with P4s instead of P3s or Athlons, then yes, it will effect the curve. However, I don't see that happening in any major way until after RC5 is over, so I think we're in the clear.

 

[BurntKooshie]

No, the P4 only has 8 general purpose registers (these are the elemental units of function which manipulate the data). This isn't enough to bitslice RC5-64. While the P4 (and all modern x86 processors) now have renaming registers to try and alleviate this, the compiler can only &quot;see&quot; 8 registers, and that's simply not enough (if the compiler &quot;saw&quot; more than that, it wouldn't be x86 backwards compatible!).

The PowerPC specification calls for 32 general purpose registers (which also makes it slightly less sensitive to cache bandwidth). I don't know exactly how many general purpose registers RC5-64 needs, but obviously it's less than 32, but more than 8 (this assumes the presence of 128bit integer SIMD).