Message boards :
Number crunching :
How this is possible?
Message board moderation
Author | Message |
---|---|
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
How this is possible? a PII is faster than a P4 3.4 or P4 3.0?? See and examine WU ID:82841 (-no offense against you: Ken Putnam!) any explaination are wellcommed littleBouncer [EDIT}Maybe that's why it is invalid...[/EDIT] |
Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
DAMN! Your pc has a sucky integer speed. Are you sure it's a Pentium? Maybe it's a cheap knockoff and is a 'Rentium' instead of a Pentium. Behold my celeron! 2.2 ghz. Measured floating point speed 1095.72 million ops/sec Measured integer speed 3394.65 million ops/sec compared to your lame 'Pentium' 3.4 ghz Measured floating point speed 1432.76 million ops/sec Measured integer speed 1818.02 million ops/sec Edit: I'm guessing Mr. P's pc had an error and exited the calculation, but exited with result code 0. I'm not the LHC Alex. Just a number cruncher like everyone else here. |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
> DAMN! Your pc has a sucky integer speed. Are you sure it's a Pentium? > Maybe it's a cheap knockoff and is a 'Rentium' instead of a > Pentium. > > Behold my celeron! 2.2 ghz. > Measured floating point speed 1095.72 million ops/sec > Measured integer speed 3394.65 million ops/sec > > compared to your lame 'Pentium' 3.4 ghz > Measured floating point speed 1432.76 million ops/sec > Measured integer speed 1818.02 million ops/sec > > > Edit: > I'm guessing Mr. P's pc had an error and exited the calculation, but exited > with result code 0. I think too. > > @ Alex That is the effect of the CC 4.25 with CC 4.19 the 'integer speed' is 2700 to 2800 (for cheating!) ('floating point speed' is equal.) (even my Notebook is faster per WU, but the 3.4 is an HT and crunches two at the time.) greetz littleBouncer |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
[QUOTE from you] >> your lame 'Pentium' 3.4 ghz [/QUOTE] That's why this Host is 14.th in RAC Ranking! 1,254 14 GenuineIntel Intel(R) Pentium(R) 4 CPU 3.40GHz 2 Windows XP 2,738.76 92.51 Or see here greetz littleBouncer |
Send message Joined: 2 Sep 04 Posts: 378 Credit: 10,765 RAC: 0 |
Heh. :) I'm not the LHC Alex. Just a number cruncher like everyone else here. |
Send message Joined: 27 Sep 04 Posts: 102 Credit: 7,340,888 RAC: 6,218 |
> How this is possible? > > a PII is faster than a P4 3.4 or P4 3.0?? > > See and examine WU > ID:82841 > > (-no offense against you: Ken Putnam!) > > any explaination are wellcommed > littleBouncer > [EDIT}Maybe that's why it is invalid...[/EDIT] > > Don't know why the completion time is so low, but the machine is actually a PIII/450, not a PII I'll have to take a look at it when I get home. (If you look at the other results, they are all closer to what would be expected) |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
> Don't know why the completion time is so low, but the machine is actually a > PIII/450, not a PII I'll have to take a look at it when I get home. > > (If you look at the other results, they are all closer to what would be > expected) > > > @ Ken Putnam when I opened this thread there was only your result and from another, so I wasn't granted (I was the 3th incoming-result). Later I saw my results and others are granted, only yours not. So I thought:"Maybe your 'sim'WU didn't make the full turn (some particles have left the colider or are colided), that's why it was 'invalid'; and once again I was too fast with my questions..." Sorry no offense against you respectfull littleBouncer |
Send message Joined: 23 Oct 04 Posts: 358 Credit: 1,439,205 RAC: 0 |
A Question: Why does it(LHC-appl. 4.64) run only on one instance on a 'two-instances-host'? With the new 'stock' from 4.4.05 I noticed; that I have to update manually for getting to run the second instance after the Client has finished a WU(before it was working as it should). BTW:The prefs' are set to use 2 CPU's. greetz littleBouncer |
Send message Joined: 17 Sep 04 Posts: 190 Credit: 649,637 RAC: 0 |
this behaviour is not new. it can be observed on a various number of boinc projects and many users reported this of several forum. Me too having this from time to time, but not in way that I should worry about. first was thingking, with the 4.25 clients this is fixed, it looks like it's not. From the obsevation, it looks like, when the client, is running especially on HT (mp?) hosts, AND having a lot of pausing/restarting switching, there is a higher risk to get a non swiching back and only one instance is seen in the memory. But I don't have an evidence for that take a look into the taskmanager, if there are one or two visible instances of the project (LHC :sixtrax...) AND if both uses cpu/memory. In my case, the restart of the boinc client solved the problem and with fresh memory allocations I played with the remove from memory preference yes/no, could not see a difference. (it also happens as mentioned, very irregulary and not all day) visite other forum, you will see, this is stil an unanswered Q.. |
©2024 CERN