Question on P-1

[JuanTutors]

Does Primenet save the P-1 residue? If so, say the person wh ran it first alotted 8MB ram, and I am alotting it 500MB ram. Will my computer start the P-1 test where the other one left off? (If not, that would be a NICE feature.)

[geoff]

It might be nice but for the current exponents being tested each residue is about 3MB in size.

[S00113]

Originally Posted by geoff

It might be nice but for the current exponents being tested each residue is about 3MB in size.

That's not much theese days. Too much for Primenet, but not much for one person or a team with a few powerful machines. As an option it would be useful if you have several machines with little RAM and a few with a lot. The less powerful machines can do step 1 and some of step 2 until it becomes more effective to continue on one of the machines with a lot of RAM.

You may do only stage 1 on a machine saving the residue by setting in prime.ini:
[pre]
SequentialWorkToDo=0
OnlyRunStage2WithMaxMemory=1
[/pre]
local.ini (set NightMemory to the amount availiable on the powerful machine):
[pre]
DayMemory=128
NightMemory=2048
DayStartTime=0
DayEndTime=0
[/pre]

Then move m####### and the corresponding line in worktodo.ini to a machine with more RAM when step 1 is done. Perhaps in this case mprime should continue doing step 2 as far as effective with 128MB RAM, and do the GCD so far before saving.

[JuanTutors]

Would it be possible to network Prime95 the same way (gasp) KaZaa or Napster used to do?

[cheesehead]

Originally Posted by dominicanpapi82

Would it be possible to network Prime95 the same way (gasp) KaZaa or Napster used to do?

IMO that's not a good idea. We need the centralization for quality control.

And I've heard that KaZaa is notorious for spreading spyware or adware -- since I'm not familiar with KaZaa and have never used it, I don't know if that reputation is actually deserved, but we certainly do not want to implement anything that has any chance of propogating malware.

[ColdFury]

And I've heard that KaZaa is notorious for spreading spyware or adware -- since I'm not familiar with KaZaa and have never used it, I don't know if that reputation is actually deserved, but we certainly do not want to implement anything that has any chance of propogating malware.

That's caused by the users on the network, not the protocol itself.

[cheesehead]

Originally Posted by ColdFury

That's caused by the users on the network, not the protocol itself.

So, you're saying the KaZaa protocol does allow transmission of malware, in contrast to current PrimeNet protocol, which doesn't. That was my point, right?

[ColdFury]

Originally Posted by cheesehead

So, you're saying the KaZaa protocol does allow transmission of malware, in contrast to current PrimeNet protocol, which doesn't. That was my point, right?

Well if primenet did use a KaZaa style protocol (and I don't see why it would, as it makes no sense for primenet's requirements) it wouldn't allow malware to be spread because it wouldn't be distributing files.

[JuanTutors]

I really shouldn't have written "the same way kazaa...". I really only meant if Prime95 can be networked. The thought was just to get more factoring done or to have one machine do one type of work and immediately pass it to another when it's done. There really isn't much point in that though.

[cheesehead]

Originally Posted by ColdFury

Well if primenet did use a KaZaa style protocol

... which I did not mean to suggest ...

(and I don't see why it would, as it makes no sense for primenet's requirements)

Let me clarify that when I commented on the proposal for distributing P-1 save files, I didn't assume that such distribution would necessarily be through PrimeNet, despite my contrasting of KaZaa protocol to "current PrimeNet protocol". It could be done independently of PrimeNet.

it wouldn't allow malware to be spread because it wouldn't be distributing files.

But that (distributing files) is what is proposed earlier in this thread: distributing P-1 save files.