|
When using this resource, please read FragKrag's opening post. The Tech Support forum regulars have helped create countless of desktop systems without any compensation. The least you can do is provide all of the information required for them to help you properly. |
NVM, got my answer in the enthusiast thread.
+ Show Spoiler +I need help with buying a headset. I have specific constraints and I don't know how to choose to satisfy them. 1. It needs to isolate me from surrounding noise very well. Especially from keyboard clicks. I'm a Software Developer and I have a colleague sitting close to me. His keyboard is absolutely crazy noisy and his typing is chaotic. I really really need a solution to not hear that when I'm trying to focus. One solution for 1. is to have the music loud which kinda works with the appropriate music I sometimes listen (heavy metal and stuff). But sometimes I want to listen to less loud music. 2. It needs to not be heard loudly by the other colleagues. I have this older colleague, sitting a little further, that doesn't like to hear loud music coming from my headset (which happens when I up the volume for heavy metal to block out the other colleague) 3. My budget goes as high as a Steelseries Siberia v.2 but I want to cover very well the two constraints, especially the first one, and however good SSv2 is, I don't know whether it is appropriate. + Show Spoiler +
Each table is 1.2 m x 0.5 m. The distances typed in are between tables.
Wall - 0.5m - empty - 0.5m - Loud Typist - 0.5m - Sensitive Old Guy - 0.5m - Wall/Window
|
0.5 m
|
Wall - 0.5m -MYSELF-0.5m -Empty Table- 0.5m - Empty Table ------- Door - Wall/Window
|
|
|
NVM, got my answer in the enthusiast thread.
+ Show Spoiler +I asked in some headset thread, then a mod close it and suggested other headset thread. So I posted there and nobody said a word on that thread for several days now. So I'll wait a bit here and follow your suggestion if nobody can help me here.
|
On February 13 2012 08:20 Kerans wrote:+ Show Spoiler +Hi guys, I submitted my build on this awesome thread a few months ago and finally built my first rig from the ground up. I recently dabbled in overclocking my i7-2600k, and since the mobo (Asus P8Z68 Deluxe/Gen3) has this little utility called TurboV EVO, I gave it a spin before going into the BIOS and starting to mess around manually.
The case is an NXZT Phantom with every possible fan except the 200mm side fan, since it wouldn’t fit together with the Cooler Master Hyper 212 HSF, and room temperature is around 18 degrees Celsius. Speed of intake fans is close to minimum, a little higher for exhaust fans - since I seem to recall that it should be the ideal setting in order to achieve maximum cooling - and the HSF blows air directly towards the back 140mm fan.
I started with the “fast” setting, and upon rebooting, the multiplier was set to x43, with a CPU voltage of 1.195, BLCK at 100 and DDR voltage at 1.5 (VCCSA Voltage: 0.925, VCCIO Voltage: 1.05, CPU PLL Voltage: 1.80, PCH Voltage: 1.05, all data comes from what I can see under the TurboV EVO section of the Asus AI Suite, under “Manual Tuning”). I trusted the thing only to a certain extent, so I got myself Prime 95 and Realtemp and kept the rig under test for close to 50 minutes. The cores’ temperatures with system idle is around 25-27 degrees, and no core went above 64-65 C during the test, with a difference in temperature between cores no wider than 3-5 degrees.
What I’m wondering is this: are these results too good to be true, or can I keep things as they are for a daily usage, which includes anything from audio/video editing to gaming? System looks very stable as far as I can tell, no problem detected whatsoever, no BSODs, no nothing - but I wouldn’t want things to come back and bite me in the ass a few months down the road.
Thank you, and thanks again to everyone who keeps this thread so useful.
A 50 minute Prime95 run isn't enough to declare the overclock 100% stable. You'll probably want to do like 20 runs of LinX with "all RAM" option (or something like that). Prime95 can be a bit slow to find errors, and I'm not sure if it will find stuff that LinX does not, but you need up to 12Hrs of Prime to be sure. I know I've found errors after 2hrs and had to boost VCore. I'd just go the LinX route and if it's stable with that, just use it normally. If you have no issues for a week, I'd consider it stable.
I've also found that 5 runs of IntelBurnTest is fast and is a good "smoke test" to find any big initial instability.
But anyway those temps are fine and if you find it stable, you're good to go. 43x is really mild anyway. I'm really disappointed you got a 2600K and a P8Z68 Deluxe/Gen3 and only a 212+ (or evo?). It makes me cringe.
|
On February 14 2012 01:24 Wabbit wrote:Show nested quote +On February 13 2012 08:20 Kerans wrote:+ Show Spoiler +Hi guys, I submitted my build on this awesome thread a few months ago and finally built my first rig from the ground up. I recently dabbled in overclocking my i7-2600k, and since the mobo (Asus P8Z68 Deluxe/Gen3) has this little utility called TurboV EVO, I gave it a spin before going into the BIOS and starting to mess around manually.
The case is an NXZT Phantom with every possible fan except the 200mm side fan, since it wouldn’t fit together with the Cooler Master Hyper 212 HSF, and room temperature is around 18 degrees Celsius. Speed of intake fans is close to minimum, a little higher for exhaust fans - since I seem to recall that it should be the ideal setting in order to achieve maximum cooling - and the HSF blows air directly towards the back 140mm fan.
I started with the “fast” setting, and upon rebooting, the multiplier was set to x43, with a CPU voltage of 1.195, BLCK at 100 and DDR voltage at 1.5 (VCCSA Voltage: 0.925, VCCIO Voltage: 1.05, CPU PLL Voltage: 1.80, PCH Voltage: 1.05, all data comes from what I can see under the TurboV EVO section of the Asus AI Suite, under “Manual Tuning”). I trusted the thing only to a certain extent, so I got myself Prime 95 and Realtemp and kept the rig under test for close to 50 minutes. The cores’ temperatures with system idle is around 25-27 degrees, and no core went above 64-65 C during the test, with a difference in temperature between cores no wider than 3-5 degrees.
What I’m wondering is this: are these results too good to be true, or can I keep things as they are for a daily usage, which includes anything from audio/video editing to gaming? System looks very stable as far as I can tell, no problem detected whatsoever, no BSODs, no nothing - but I wouldn’t want things to come back and bite me in the ass a few months down the road.
Thank you, and thanks again to everyone who keeps this thread so useful. A 50 minute Prime95 run isn't enough to declare the overclock 100% stable. You'll probably want to do like 20 runs of LinX with "all RAM" option (or something like that). Prime95 can be a bit slow to find errors, and I'm not sure if it will find stuff that LinX does not, but you need up to 12Hrs of Prime to be sure. I know I've found errors after 2hrs and had to boost VCore. I'd just go the LinX route and if it's stable with that, just use it normally. If you have no issues for a week, I'd consider it stable. I've also found that 5 runs of IntelBurnTest is fast and is a good "smoke test" to find any big initial instability.
Keep in mind that LinX and IntelBurnTest are basically the same thing. They're both frontends for LinPack, which is a library for a whole bunch of cool linear algebra routines. So both programs just do tons of matrix inversions and the likes while showing you a nice little GUI window.
|
Ah, didn't know that. I've just always used IBT but I have read that a lot of people on overclock.net and other places just use LinX.
|
Thanks Wabbit & Rannasha, I'll proceed with Linx and let you know. Btw, is the CM 212 cheap-ish for this kind of rig? I guess I could always upgrade it sometime down the road, any suggestions?
|
You have an ASUS Deluxe so yes it is.
|
I wouldn't bother taking a Hyper 212 out to replace it unless you somehow get an urge to overclock a lot more without increasing the noise from how it is now, or maybe you get a new motherboard. Higher-end heatsinks exist because some people want them, and they're more necessary for cooling processors that use a lot more power. If it works fine for you, it works fine.
On a side note, do people usually use language like "find errors" when stress testing overclocks? Somehow this is conjuring the wrong mental image, like one program searches the CPU for errors quicker than another. Maybe it's just me. Under the surface it's really just that under certain workloads with current flowing from one place to another, given the power supplied through the electrical grid through the power supply through the motherboard VRMs and some unlucky amount of noise, you eventually get a voltage somewhere that more or less is interpreted as a 0 even though it was supposed to be a 1, or vice versa. That's the error. For stability you just want the probability of error under all reasonable workloads to be vanishingly small.
edit: not saying that anything is wrong or different words should be used, just an observation. It really depends on your perspective.
edit2: while I've got the soapbox out, I don't know wtf some people think a PWM, PLL, or spread spectrum are heh. That's an unrelated point.
|
Oh shit I get to use my computer engineering resources here.
Yes Myrmidon that is one case where voltage controls will cause a CPU to give a logic error, this is what happens when too much voltage is applied to logic gates and circuitry. Simply put, if given boolean algebra of AB=X, X is only supposed to equal 1 where both A and B are 1. However the reality is that if even A or B (but not both) is a logic 1, some current still passes through the gate. For the sake of argument lets say about 15 percent of voltage bleeds through. In a 1 volt chip, if you supply an input A of 10 volts of potential difference and a B input of 0. Even though this is a logic zero, the X value will actually be a logic 1 because the component was brute forced by input A.
Additionally, we increase voltage to propagate electricity through the system faster (in computer "bullet time" elements dont react instantly, they slowly "warm up" over time, although not a lot of time lol). Because if you toggle a processor too fast, you can actually run into the problem where you are making a new clock cycle before the previous one has had a chance to completely run its course through the processor.
Which isnt necessarily a bad thing until you get to state machines, and a complex processor where certain paths to the end take way more time than others.
Thats why our objective is generally to run as high a clock speed as we can using minimal voltage necessary to keep the processor from logically folding into itself.
Its that last part where different tests work better than others. A good test acknowledges the architectural weaknesses of the processor it is running on and will hammer slow areas of a processor and fast areas in as close proximity as possible. If a processor can handle this brutal workload and come out the other side fine then it is as seen as having no "errors."
Though I agree that "finding errors" is a bit of a misnomer. Its more of a physics test really. The error is the ultimate result, not the cause.
EDIT: I know you already know all this. I guess Im just expounding on why the right test works better and for others that dont know wtf is going on.
|
looking for a quality z68/p67 mobo thats able to run crossfirex. im limited to Microcenter @ CA OC/Tustin because of their awesome -50$ off a p67/z68 mobo with a purchase of a 2500k.
|
Brunei Darussalam760 Posts
can i play max settings with a amd phenom ii x4 955, powercolorhd 6790 and 4gb ram??
|
Kind of, though the professor is going to cap it and you will drop in framerates in large engagements
|
On February 14 2012 18:23 iKill[ShocK] wrote:
looking for a quality z68/p67 mobo thats able to run crossfirex. im limited to Microcenter @ CA OC/Tustin because of their awesome -50$ off a p67/z68 mobo with a purchase of a 2500k.
Just ask them at the store? They only have like 3-4 different mobos to choose from that qualify for the deal, I doubt there's more than 2 (if that) that can do crossfire. Their reps, while not necessarily the most knowledgeable, should be decent. At least, that's how it was when I went like a month and a half ago.
|
On February 14 2012 01:24 Wabbit wrote:Show nested quote +On February 13 2012 08:20 Kerans wrote:+ Show Spoiler +Hi guys, I submitted my build on this awesome thread a few months ago and finally built my first rig from the ground up. I recently dabbled in overclocking my i7-2600k, and since the mobo (Asus P8Z68 Deluxe/Gen3) has this little utility called TurboV EVO, I gave it a spin before going into the BIOS and starting to mess around manually.
The case is an NXZT Phantom with every possible fan except the 200mm side fan, since it wouldn’t fit together with the Cooler Master Hyper 212 HSF, and room temperature is around 18 degrees Celsius. Speed of intake fans is close to minimum, a little higher for exhaust fans - since I seem to recall that it should be the ideal setting in order to achieve maximum cooling - and the HSF blows air directly towards the back 140mm fan.
I started with the “fast” setting, and upon rebooting, the multiplier was set to x43, with a CPU voltage of 1.195, BLCK at 100 and DDR voltage at 1.5 (VCCSA Voltage: 0.925, VCCIO Voltage: 1.05, CPU PLL Voltage: 1.80, PCH Voltage: 1.05, all data comes from what I can see under the TurboV EVO section of the Asus AI Suite, under “Manual Tuning”). I trusted the thing only to a certain extent, so I got myself Prime 95 and Realtemp and kept the rig under test for close to 50 minutes. The cores’ temperatures with system idle is around 25-27 degrees, and no core went above 64-65 C during the test, with a difference in temperature between cores no wider than 3-5 degrees.
What I’m wondering is this: are these results too good to be true, or can I keep things as they are for a daily usage, which includes anything from audio/video editing to gaming? System looks very stable as far as I can tell, no problem detected whatsoever, no BSODs, no nothing - but I wouldn’t want things to come back and bite me in the ass a few months down the road.
Thank you, and thanks again to everyone who keeps this thread so useful. A 50 minute Prime95 run isn't enough to declare the overclock 100% stable. You'll probably want to do like 20 runs of LinX with "all RAM" option (or something like that). Prime95 can be a bit slow to find errors, and I'm not sure if it will find stuff that LinX does not, but you need up to 12Hrs of Prime to be sure. I know I've found errors after 2hrs and had to boost VCore. I'd just go the LinX route and if it's stable with that, just use it normally. If you have no issues for a week, I'd consider it stable. I've also found that 5 runs of IntelBurnTest is fast and is a good "smoke test" to find any big initial instability. But anyway those temps are fine and if you find it stable, you're good to go. 43x is really mild anyway. I'm really disappointed you got a 2600K and a P8Z68 Deluxe/Gen3 and only a 212+ (or evo?). It makes me cringe.
Just a quick update:
- IBT went smoothly for the 5 cycles you suggested, max temps were 58/61/66/61 on the cores, core voltage oscillated from 1.272 to 1.280.
- Left Prime95 to do its thing for 13 hours, everything ok, same temps and core temp differences as IBT (+/- 2 degrees), core voltage went from 1.264 min to 1.280 max.
- LinX, though, stopped at the 16th cycle of full RAM testing (each went on for 280secs give or take), after 1h 24m 3s. Max temps were 57/60/62/59.
Care to give any further suggestions?
|
Add a little bit of voltage. You should be able to get a fully stable system that way. Temperatures are still low enough to allow for that. Anything sub-70 in LinX/IBT is more than fine, day-to-day usage doesn't stress the CPU nearly as much as these stress-testers and the Sandy Bridge CPUs can work at 70 degrees just fine.
|
|
|
On February 14 2012 22:08 Kerans wrote:Show nested quote +On February 14 2012 01:24 Wabbit wrote:On February 13 2012 08:20 Kerans wrote:+ Show Spoiler +Hi guys, I submitted my build on this awesome thread a few months ago and finally built my first rig from the ground up. I recently dabbled in overclocking my i7-2600k, and since the mobo (Asus P8Z68 Deluxe/Gen3) has this little utility called TurboV EVO, I gave it a spin before going into the BIOS and starting to mess around manually.
The case is an NXZT Phantom with every possible fan except the 200mm side fan, since it wouldn’t fit together with the Cooler Master Hyper 212 HSF, and room temperature is around 18 degrees Celsius. Speed of intake fans is close to minimum, a little higher for exhaust fans - since I seem to recall that it should be the ideal setting in order to achieve maximum cooling - and the HSF blows air directly towards the back 140mm fan.
I started with the “fast” setting, and upon rebooting, the multiplier was set to x43, with a CPU voltage of 1.195, BLCK at 100 and DDR voltage at 1.5 (VCCSA Voltage: 0.925, VCCIO Voltage: 1.05, CPU PLL Voltage: 1.80, PCH Voltage: 1.05, all data comes from what I can see under the TurboV EVO section of the Asus AI Suite, under “Manual Tuning”). I trusted the thing only to a certain extent, so I got myself Prime 95 and Realtemp and kept the rig under test for close to 50 minutes. The cores’ temperatures with system idle is around 25-27 degrees, and no core went above 64-65 C during the test, with a difference in temperature between cores no wider than 3-5 degrees.
What I’m wondering is this: are these results too good to be true, or can I keep things as they are for a daily usage, which includes anything from audio/video editing to gaming? System looks very stable as far as I can tell, no problem detected whatsoever, no BSODs, no nothing - but I wouldn’t want things to come back and bite me in the ass a few months down the road.
Thank you, and thanks again to everyone who keeps this thread so useful. A 50 minute Prime95 run isn't enough to declare the overclock 100% stable. You'll probably want to do like 20 runs of LinX with "all RAM" option (or something like that). Prime95 can be a bit slow to find errors, and I'm not sure if it will find stuff that LinX does not, but you need up to 12Hrs of Prime to be sure. I know I've found errors after 2hrs and had to boost VCore. I'd just go the LinX route and if it's stable with that, just use it normally. If you have no issues for a week, I'd consider it stable. I've also found that 5 runs of IntelBurnTest is fast and is a good "smoke test" to find any big initial instability. But anyway those temps are fine and if you find it stable, you're good to go. 43x is really mild anyway. I'm really disappointed you got a 2600K and a P8Z68 Deluxe/Gen3 and only a 212+ (or evo?). It makes me cringe. Just a quick update: - IBT went smoothly for the 5 cycles you suggested, max temps were 58/61/66/61 on the cores, core voltage oscillated from 1.272 to 1.280. - Left Prime95 to do its thing for 13 hours, everything ok, same temps and core temp differences as IBT (+/- 2 degrees), core voltage went from 1.264 min to 1.280 max. - LinX, though, stopped at the 16th cycle of full RAM testing (each went on for 280secs give or take), after 1h 24m 3s. Max temps were 57/60/62/59. Care to give any further suggestions?
OCCT is a pretty good tool to stresstest which is often overlooked due to the popularity of prime95. They have their own stresstest, and linpack. Pretty good for general stresstesting too (since it includes PSU/GPU), and includes monitoring.
|
Has anyone heard of the i5 2550k? Is it better than the i5 2500k? More overclockable? Any FPS/other benchmarks?
|
it just has no integrated graphics i think
|
|
|
|
|
|
EPT
