Q6600 @ stock

[narreth]

How much points per day does a Q6600 @ 2.4GHz get on average in BOINC?

 

[Fardringle]

It depends on the project. Some give a lot more points than others for the same amount of work.

 

[narreth]

say im running 2 cores Climate prediction and 2 cores rosetta?

 

[petrusbroder]

Here are some numbers for all boinc-projects. This does depict just the average over all BOINC-projects.

Using the project list at the left side you can easily look up the points for each project and compare the CPUs.

Be aware of the limitation that quite a few crunchers do as you do --- crunch several projects at the same time and that thus the numbers are highly variable depending which projects are mixed: If you e.g. combine ABC@Home and yoyo@home (these project are very generous with credits) you'll get much higher points then if you combine WorldCommunityGrid and MalariaControl (where the credits are much more conservatively distributed).

I would probably trust the seti@home numbers most because there the number of FLOPS really determines the credits given. OTOH the numbers are still a mix of the old way of calculating the credits with the new (FLOPS) ...

 

[bryanW1995]

peter that list is misleading imho. it shows the q9450 averages over 1k ppd per core, while a q6600 is only 536, and a q6700 is 330!!! also, 4000 per core for a cpu is obviously in a VERY high-credit project.

 

[petrusbroder]

Originally posted by: bryanW1995
peter that list is misleading imho. it shows the q9450 averages over 1k ppd per core, while a q6600 is only 536, and a q6700 is 330!!! also, 4000 per core for a cpu is obviously in a VERY high-credit project.

I agree - but is there a better way unless you check the efficiency project by project?
If you look you'll see that you can check the efficiency of different CPUs for each and every project that BOINC-stats publishes stats on - and then you can compare.

OTOH: if there are only a few CPUs of each kind registered in one project the numbers become very biased too ...

There is essentially no good way to find out - unless you do it yourself. And that is hard work. I did that some 3 years ago for my CPUs then - the thread is somewhere.

If you collect data you have to take into account so many things (beside the type of CPU): RAM, FSB, OC:ed?, CPU-temp (because if the temps are too high the CPU throttles down) what version of BOINC?, what projects are run at the same time, what way are the credits calculated (a quorum of two? of three? of four and the extremes are dropped? FLOPS? set credits for each type of WU?), if there are plenty of WUs available, if you crunch 24/7 or part time?, if you use the comp for other stuff too?, etc, etc, etc.

These variables cancel each other if the numbers of CPUs is sufficiently high and if the project is sufficiently stable. There are only 3 - 4 projects which fullfill the criteria IMHO: seti@home, einstein@home, rosetta@home, come to mind, Folding@home (but that is not BOINC), many of the others are too small: QMC@Home has - for example - only 16 CPUs of the type Intel Q9450 @ 3.00GHz ... How do compare that processor's numbers with the 1 238 CPUs of the type Q6600 @ 2.40GHz?

I have really no idea how to compare the CPUs in a fair way. Perhaps there is no good way ... 🙁
OTOH: It is a lot of fun to get a new CPU and try it out ... 😉

 

[bryanW1995]

I believe that the best way for us to compare is for intel and amd to ship us every c2d/phenom chip they offer with appropriate mobo/case etc and good aftermarket cooling. then we can determine for ourselves 🙂

 

[dajeepster]

Originally posted by: bryanW1995
I believe that the best way for us to compare is for intel and amd to ship us every c2d/phenom chip they offer with appropriate mobo/case etc and good aftermarket cooling. then we can determine for ourselves 🙂

I second that suggestion

 

[narreth]

no one on the TeAm has a q6600 running stock?
>_>

anyway i got the q6600, this is so much faster than my sempron 2800+ -_-;;

 

[bryanW1995]

I was getting 1100 ppd per core on an e6750 @3.4 using an optimized v2.4 seti@home app for 9 mos. the optimization added about 25% to my score, so drop that down to 825 ppd per core. then 825 * 2.4/3.4 ~ 582 per core per day on seti. that seems to jive well with petrusbroder's list that shows an avg of 536 ppd, esp when you factor in that my ram was at 5-5-5-18 ddr2 1062 and I had a trd of 6. I would use the 536 per core number as a good average, so a little over 2100 ppd.


btw, that chart isn't quite so good for x3350. there are only 88 x3350 "cores" in the database, so only 22 total rigs. My unit is getting 1300ppd per core right now and is climbing fast. the x3350 (same as q9450) is significantly ahead of q9450, 9550, and 9650. However, as peter mentioned, a low sample number could easily skew the results, esp if many of those were highly overclocked cpus.

 

[narreth]

how do i set boinc to utilize all 4 cores?
Do i just open up 4 boinc managers and it sorts itself out? or is there something more i need to do>

 

[Philippart]

in the advanced view, click the advanced menu, preferences and set 4 CPUs at the bottom (or 100% in newer alpha/beta clients)

 

[Assimilator1]

BOINC automatically detects & uses all cores AFAIK.

Originally posted by: narreth
no one on the TeAm has a q6600 running stock?
>_>

anyway i got the q6600, this is so much faster than my sempron 2800+ -_-;;

Awesome! 😀, but why would you want to run it at stock?😕
Mine's at 3GHz 🙂.

 

[narreth]

i never oc'd before and im pretty happy with the speed @ stock. I plan to OC in the future though.

@ stock the q6600 is 8x as fast as my sempron 2800+ (at crunching)

Load temps are 60C, 57C, 59C, 60C

these are OK temps right? It's on an arctic freezer pro

 

[Philippart]

yes, those temps are ok, but I would buy a new cooler before overclocking

 

[bryanW1995]

you need a better cooler to get some SERIOUS overclocking done, but if you just want a small speed boost you should be able to get to 3.0 ghz without changing anything other than fsb. You know you want to try it...your hand is moving towards the f2 key...go to the bios menu...change fsb to 333...reboot... It really IS that simple on most G0 quads, by the way.

 

[Assimilator1]

For most mbrds it's the 'delete' key 😛😉.

I needed to take my G0 Q6600 from 1.25v to 1.3v to get 3GHz, YMMV.

Originally posted by: Philippart
yes, those temps are ok, but I would buy a new cooler before overclocking

Yea roger that :thumbsup:.
Though he could just go up in small steps & see what the temp does.

Narreth
Nice getting a massive speed boost like that isn't it?😀, I'm not far off that too, compared to my Ath XPM @2.5 GHz I had upto a year ago anyway.

Btw re overclocking ,let's say that you were to go for a modest overclock of 3GHz in 1 yrs time, how many credits will you of lost by not doing that 600MHz boost tomorrow?😉 (or whenever you get a better cooler, if need be).

 

[narreth]

yes it is so nice to get a massive speed boost. Since im an uneployed 17 year old, I can only afford to upgrade every 4 years or so. That's the main reason I went with the Q6600 (over a E8400)!
I will overclock soon though! 🙂

but now we have +2000 ppd for the teAm!

 

[bryanW1995]

I'm telling you, overclocking is so much fun!!

what are your load temps right now with coretemp and realtemp? delta to tjunction?

 

[Assimilator1]

He quoted his load temps earlier, though I must admit I forgot to ask what app he was using 😱.

narreth
Yea in the meantime 2000ppd is a great boost for the team 😀.