-
Hey, guest user. Hope you're enjoying NeoGAF! Have you considered registering for an account? Come join us and add your take to the daily discourse.
Race Your PC Thread of Benchmarking to find out who has the fastest PC on GAF!
- Thread starter ColonialRaptor
- Start date
Yeah, you are right about volts/heat, I wasn't aware that the modern CPUs could actually get that high, seems absurd.gunbo13 said:
I know some people pumping volts into the first gen i5/i7 (and ICH) and I believe some of the second gen i5/i7 had problems. Along with the fact people are running 32nm chips at the same voltages as 45nm is what makes me really hesitant about upping the volts even thought temps are in check.
So basically it is just me being conservative.
Just want to update my score slightly. New BF3 drivers bumped me up a bit. Same specs, but score and link below:
P8323 http://3dmark.com/3dm11/1898221
P8323 http://3dmark.com/3dm11/1898221
Vulcano's assistant
Banned
Well, I finally made it run at 845 Mhz, as I suspected it was lacking voltage. This is my first successful test in the entire night so it is kind of rough in the edges. And I can't say I am happy with the performance.
P5495 - 1 x Asus Nvidia GTX 570 1280 MB - 845 Mhz / 2150 Mhz / 1.1 V - i5 2500k @ 5 Ghz - 8GB - Asus p8p67 PRO - Vulcano's assistant - Air - Link
P5495 - 1 x Asus Nvidia GTX 570 1280 MB - 845 Mhz / 2150 Mhz / 1.1 V - i5 2500k @ 5 Ghz - 8GB - Asus p8p67 PRO - Vulcano's assistant - Air - Link
ss_lemonade
Member
ColonialRaptor
Member
Hope you guys don't mind but I'm taking a day off updating the ladder today, I've spent 2 - 3 hours updating it each day since I started the thread and it's dragging on me... I'll update it again tomorrow or the day after, my attention needs to be on getting a BF3 code at the moment...
gunbo13 said:
I wouldn't go so far as saying P95 is retired, linx flags instability way sooner though but I've had errors flag with prime blends 6 to 8 hours in while my rig was able to do 50 maxed linx passes without issues. Also, linpack generates tons of heat, if you're cooling is not up to it long P95 sessions are the way to fly. (My personal way of checking stability is 20maxed linx runs + 12 hours P95 blends, but everyone has different opinions on that :lol)
OCN really was better a couple years back, but that could just be me feeling nostalgic. Whatever. Socket 775 overclocking was so well documented on there, socket 1366 as well. Can't say the same for SB related stuff though, but that might be because I never really bothered to look since I'm still on 1366. Anyway, there are still some gurus left in the motherboard specific threads but it feels like it's slipped a bit yeah. Might also have something to do with these new mobos and chips, everyone can OC now. On socket 775 you still needed to fool around with clock skewing and clock drive, the real OC voodoo, now it all looks/feels easy.
/rant
ColonialRaptor said:
Must be a royal pain in the ass to change every ranking number manually
What about setting up some kind of google spreadsheet thing? See that a lot on OCN. Probably easier to manage.example
Is there one for 6950 or 69xx AMD cards? I poke around but didn't see anything besides owners thread.koji said:I wouldn't go so far as saying P95 is retired, linx flags instability way sooner though but I've had errors flag with prime blends 6 to 8 hours in while my rig was able to do 50 maxed linx passes without issues. Also, linpack generates tons of heat, if you're cooling is not up to it long P95 sessions are the way to fly. (My personal way of checking stability is 20maxed linx runs + 12 hours P95 blends, but everyone has different opinions on that :lol)
OCN really was better a couple years back, but that could just be me feeling nostalgic. Whatever. Socket 775 overclocking was so well documented on there, socket 1366 as well. Can't say the same for SB related stuff though, but that might be because I never really bothered to look since I'm still on 1366. Anyway, there are still some gurus left in the motherboard specific threads but it feels like it's slipped a bit yeah. Might also have something to do with these new mobos and chips, everyone can OC now. On socket 775 you still needed to fool around with clock skewing and clock drive, the real OC voodoo, now it all looks/feels easy.
/rant
Must be a royal pain in the ass to change every ranking number manually
example
And why not? These latest processors are being overclocked to ridiculous levels without a steep learning curve. Most high clocks are just chasing numbers with really no practicality to 99% of users. In past generations, 100% overclocks were used to equal top of the line processor options or to achieve practical results. Now, 100% overclocks are for competition yet the idea still leaks to end-users. Which I believe is a mistake. You should be conservative since your gain is about zero and that is the right call. My 920 is a fantastic chip and I certainly have a history with overclocking. My current voltage? Stock.Hazaro said:
You just don't have to endanger your CPUs these days for 24/7 use. Now your video cards...those things should burn.
It's not retired but that doesn't really mean it is that useful anymore. Typically if linx passes you are not failing p95 on CPU stability but MB/RAM. I don't condemn the test because the more the merrier. But most overclockers I worked with graduated from p95 years ago.koji said:
More heat means more stress. Avoiding that by using p95 is not the right call. If your cooling is not up to par with linpack, you have to step it up or dial back. Dodging it just means you are not achieving a proper level of stability.
gunbo13 said:
Yeah that's what I do myself, linx to check if my CPU OC is stable then I move to long blends sessions, if something borks up there it's probably VTT/mem.
You're right about the heat, it's an option though, not a lot of air-cooled systems can dissipate the heat generated by hours of linpack on a seriously OC'ed CPU.
lowrider007
Licorice-flavoured booze?
P6311 - 1 x Palit GTX 570 (c850Mhz/s1700/m2000Mhz 1.050v) i5-2500k @ 4.5Ghz 3.00v - Cosair 8GB DDR3 1600Mhz - Asus P8Z68-V - Enermax Liberty 620w - lowrider007 - air - Link
Managed to improve on my score a fair bit, this is quite fun tbh, never o/c'd my GPU before, this thread has certainly got me learning how to push my system, later on I may attempt to push my gpu core to 900Mhz and my CPU to 4.8Ghz, I'm still in the trenches but I feel that my system is getting 'comparable' results now at least.
Managed to improve on my score a fair bit, this is quite fun tbh, never o/c'd my GPU before, this thread has certainly got me learning how to push my system, later on I may attempt to push my gpu core to 900Mhz and my CPU to 4.8Ghz, I'm still in the trenches but I feel that my system is getting 'comparable' results now at least.
Well I don't want to get into a big discussion here but if an air-cooled system can't dissipate the heat generated by linpack, it should not be clocked that high. These software packages don't cause higher thermal output by chance. One is stressing the CPU in an entirely different way, typically with more heat meaning more stress. So the only "option" in that regard using another stress test like p95, is to result in p95 stability. That doesn't necessarily mean your system is stable, handling all processing demands. It only means it is stable regarding a p95 test, which IMO is not adequate enough. I typically show a tiny grimace each time I see p95 stable for x amount of hours since I can envision the future problems for users.koji said:
However, if you don't see it, it doesn't exist, is common for overclocking. This means that a BSOD might never occur with a system that is p95 stable due to the nature of the user demand. Does that mean the system is stable for all processing? No, that would only be a game of chance. The same though applies for all stress tests as they are simply simulations, with some being more practical then others. And even if a p95 achiever gets a few BSOD's they will probably chalk it up to a video card crash or something. They will then focus their mindset back to their machine being stable citing the p95 testing and this is just the nature of computing.
My opinion is that you shouldn't implore bad practices on a testing methodology that is not even completely accurate to real-time events. Many overclocking practices out there are inadequate. I covered some tips ITT but true stability requires effort. If you put in the effort, you get machines which don't encounter BSOD's at all.
lowrider007
Licorice-flavoured booze?
Over the years I've really started to relax when it comes to IBT'ing or P95'ing my system for hours on end, my system is never going to be stressed like that in it's life, if anything I'm causing more harm than good, ok I may get a BSOD in the life of my machine, if that happens then I up the voltage or down the multiplier, the thing is what with CPU's down-stepping these days and with the offset voltages on the new chips your pretty much guaranteed stability for 90% of the time even without a serious bench test, with these new chips you only really have to worry about stability on high loads which for most people means gaming, and then you also have to consider that most games are GPU limited, so the gap for when you need true stability lessons and lessons for the average user.
You see the thing is with the i5's using the offset voltage and down-stepping I can have a rock solid system @ 1600Mhz (x16 multiplier) @ around 0.960v's, which is for the majority of the systems life, it's only when it jumps the multiplier and voltage up that it steps into the realms of questionable stability and even then it depends on how much load your putting on the CPU.
You see the thing is with the i5's using the offset voltage and down-stepping I can have a rock solid system @ 1600Mhz (x16 multiplier) @ around 0.960v's, which is for the majority of the systems life, it's only when it jumps the multiplier and voltage up that it steps into the realms of questionable stability and even then it depends on how much load your putting on the CPU.
Your not causing any harm but I do understand what you are saying. The pain comes from people in threads talking about how their game crashes. Many times you read a statement such as "OK, I'm going to lower the clock speed on my processor to see if that helps." Oh great it works now! But was that really the overclock and not a false positive? Why are you even doing that and how were you operating with faulty settings? It makes no sense.lowrider007 said:
The problems people experience and dislike about overclocking rarely occur early. You see in this thread that people are talking about sudden changes in configurations, etc... The majority of these claims, and almost coming from elsewhere, just comes down to the idea of not properly testing your system (no flak intended and I do say "majority"). Would you rather commit to extra effort in the beginning or scramble to fix a problem 6 months down the line? It's a choice and most people choose the latter.
That's pretty much it in a nutshell. When I have had system issues in the past I can immediately treat my overclocking settings as a last resort. And typically, the settings are not the case. It's not a bad thing to have... Yet people balance their decisions differently then myself so I accept that. I still do however try to convey the idea of proper practices even if it results in helping just a few people not having headaches 6 months down the line.
bloodydrake
Cool Smoke Luke
woober said:
I got the msi r6990 it came with msi afterburner that lets you tweak card
Peterthumpa
Member
LOL, went to format my PC only for this game and almost lost the whole rig due to a faulty SATA cable, I mean, fire and shit.
Thanks God it still boots, though I lost my blu-ray drive. Dammit!
Thanks God it still boots, though I lost my blu-ray drive. Dammit!
bloodydrake
Cool Smoke Luke
Ok got the other 6990 and setup 2X Crossfire
Heres results for Single and then 2X Crossfire
P10225 - 1x R6990 - Core@900/Mem@1400 - i7 2600k@4.4 - Kingston 8GB 1600 @ stock 1600 - Asus P8Z68 Deluxe - BloodyDrake - Air - Link
P14864 - 2x R6990 - Core@900/Mem@1400 - i7 2600k@4.4 - Kingston 8GB 1600 @ stock 1600 - Asus P8Z68 Deluxe - BloodyDrake - Air -
Link
Heres results for Single and then 2X Crossfire
P10225 - 1x R6990 - Core@900/Mem@1400 - i7 2600k@4.4 - Kingston 8GB 1600 @ stock 1600 - Asus P8Z68 Deluxe - BloodyDrake - Air - Link
P14864 - 2x R6990 - Core@900/Mem@1400 - i7 2600k@4.4 - Kingston 8GB 1600 @ stock 1600 - Asus P8Z68 Deluxe - BloodyDrake - Air -
Link
$1,400 in graphics and you still have a stock cooler? :[
Yup. GPU voltage I just cracked up to 1.3V pretty fast. Pssh, it's fine.gunbo13 said:And why not? These latest processors are being overclocked to ridiculous levels without a steep learning curve. Most high clocks are just chasing numbers with really no practicality to 99% of users. In past generations, 100% overclocks were used to equal top of the line processor options or to achieve practical results. Now, 100% overclocks are for competition yet the idea still leaks to end-users. Which I believe is a mistake. You should be conservative since your gain is about zero and that is the right call. My 920 is a fantastic chip and I certainly have a history with overclocking. My current voltage? Stock.
You just don't have to endanger your CPUs these days for 24/7 use. Now your video cards...those things should burn.
MSI Afterburner. You may have to modify your registry to get to higher clockspeeds.woober said:
bloodydrake
Cool Smoke Luke
Just waiting for the H60's(got 5 on order) to arrive. hopefully by midweekHazaro said:
lowrider007
Licorice-flavoured booze?
ColonialRaptor said:
Booooooo!!
Thanks for the work you put in, I'm always scared of starting a thread of a similar vein due to how much time it takes up, I'm not sure how many people realise the work involved in what looks like simple tables, but I know they can be quite stressful at times sorting them out.
whatsinaname
Member
P3909 - 1 x XFX 6950 2GB (800Mhz/1250Mhz/0.900v) - E8400 @ 3.03 Ghz - 2 x 2048 MB G.Skill 7 @ 400 MHz - Gigabyte EP45-UD3P - Air - Link
Damn, I am doing my 6950 a disservice by sticking to the E8400. But I play games that are 6 months - 1 year old, so don't really see a need to upgrade right now. Only reason I got a 6950 is because I got a 2560x1600 monitor.
Damn, I am doing my 6950 a disservice by sticking to the E8400. But I play games that are 6 months - 1 year old, so don't really see a need to upgrade right now. Only reason I got a 6950 is because I got a 2560x1600 monitor.
ss_lemonade
Member
OC that wolfdale chip, you should still be able to squeeze out more performance in a lot of games. I think 3d mark 11, like previous versions still gives out big points for having fast CPUs, which is why you're seeing low scores.whatsinaname said:
Score - 1x Radeon 6850 1GB 2500k @3.2 GHz 4GB ram at 1600 (I think) asus p8p67 Ultrabum Air http://3dmark.com/3dmv/3511569
So yeah, why does it say that my stuff is lower compared to similar spec computers, is it because I don't have my 2500K overclocked?
Just wondering.
So yeah, why does it say that my stuff is lower compared to similar spec computers, is it because I don't have my 2500K overclocked?
Just wondering.
lowrider007
Licorice-flavoured booze?
Ultrabum said:
It's quite common tbh, I had the same, it wasn't until I pushed my cpu to 4.5Ghz and o/c'd my video card that it said my system was comparable.
TheVampire
Banned
Comeon lets get this list updated!
alysonwheel
Member
Tried to push my creaky old C2D a little higher today, but I couldn't get it stable at 3.5. Managed to squeak out a slightly higher score, though.
P3476 - 1x nVidia GTX560 1gb (918/2000) - Intel Core 2 Duo E7200 @ 3325 Mhz - 2x 2048 MB Mushkin 7 @ 400 MHz - Gigabyte P31-DS3L - alysonwheel - air Link
edit: superseded.
alysonwheel
Member
Would you look at that!Hazaro said:
P3614 - 1x nVidia GTX560 1gb (918/2000) - Intel Core 2 Duo E7200 @ 3799 Mhz - 2x 2048 MB Mushkin 7 @ 400 Mhz - Gigabyte P31-DS3L - alysonwheel - air Link
That's probably as high as everything in my system will go. I guess I could try clocking the RAM up a bit but I think I'll wait until I've been sitting on this overclock for a few days.
Thanks for repeatedly showing me where to kick my PC, Hazaro.
DeadRockstar
Member
Decided to give this a crack on the gaming PC.
P6992 1x Palit GTX 580 (860/2010/1.000v) - i7 920 @ 3780mhz - 3 x 2GB Corsair XMS DDR @ 1333 - Gigabyte EX58-UD5 - DeadRockstar - Air - Link
Not bad with stock memory and voltage that's lower than stock on some 580s. If I had more time I'd give it another run at 870mhz and see if it was stable and over 7k score but ah well.
P6992 1x Palit GTX 580 (860/2010/1.000v) - i7 920 @ 3780mhz - 3 x 2GB Corsair XMS DDR @ 1333 - Gigabyte EX58-UD5 - DeadRockstar - Air - Link
Not bad with stock memory and voltage that's lower than stock on some 580s. If I had more time I'd give it another run at 870mhz and see if it was stable and over 7k score but ah well.
LabouredSubterfuge
Member
sn1pes said:
I don't understand this wizardry.
LabouredSubterfuge
Member
Chinner said:
Which GTX580 did you get duck?
Should be able to push the clock harder than that on it and it keep stable. Use MSI Afterburner.
LabouredSubterfuge
Member
Chinner said:
Same as me then.
With the most up to date drivers, I can push 895mhz stable as long as I turn the voltage up to 1.138v. Make sure you push your fans a little harder to keep it cooler too.
LabouredSubterfuge
Member
Chinner said:
Dude, you're not going to find too much better in the way of GPU cooling than the Twin Frozr fans besides moving to water cooling.
bloodydrake
Cool Smoke Luke
Does anyone know if having 2 pcie slots running at 8x for crossfire would slow down dual card bench?
the Mobo I'm using is an asus P8Z68 DELUXE that says slot one runs at x16 with one card but with 2 cards slot 1 and 2 both run at x8.
or is this normal behavior.?
just wondering if this is part of the bottle next in the quad configuration.
the Mobo I'm using is an asus P8Z68 DELUXE that says slot one runs at x16 with one card but with 2 cards slot 1 and 2 both run at x8.
or is this normal behavior.?
just wondering if this is part of the bottle next in the quad configuration.