You can downgrade to the i7 950 and use the extra money to upgrade to a decent SSD. They really aren't all that bad priced anymore. If you want the TB, keep it for storage but you're bottle neck is going to be the HD. Or you can get a 500gb Western Digital Passport and have portable storage. A TB is alot of space for everything you need except movies, and a TB can't hold enough movies to make it worth it, imho.
I have a very simliar system with a 1TB 7200rpm HD.
Originally posted by dfan A 600 watt psu will do, you don't need a nuclear power plant for normal computer.
I wouldn't change x58 stuff away, it's with multiple gfx cards where triple channel memory is most effective.
Wow, I actually though the 5890 had a much higher power consumption than it does. 190W for the 5890. That would have an absolute maximum power draw of 500W in an i7 920 system, so a 750W PSU would cover it easy. If he wanted to put another 5870 or 5890 into Crossfire he'd probably want a 800W-950W PSU though.
A 600W would work too but with degradation it might not quite cut it in a couple years, I like to recommend a little bit higher just to be safe so noone goes, man that guy gave horrible advice and I got screwed :P But yea 1000W are almost always uncalled for.
A 600 watt psu will do, you don't need a nuclear power plant for normal computer.
I wouldn't change x58 stuff away, it's with multiple gfx cards where triple channel memory is most effective.
Wow, I actually though the 5890 had a much higher power consumption than it does. 190W for the 5890. That would have an absolute maximum power draw of 500W in an i7 920 system, so a 750W PSU would cover it easy. If he wanted to put another 5870 or 5890 into Crossfire he'd probably want a 800W-950W PSU though.
A 600W would work too but with degradation it might not quite cut it in a couple years, I like to recommend a little bit higher just to be safe so noone goes, man that guy gave horrible advice and I got screwed :P But yea 1000W are almost always uncalled for.
Originally posted by dfan So when we reduce the psu's efficiency from total system consumption 500*0,8 = 400 watts And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
So when we reduce the psu's efficiency from total system consumption
500*0,8 = 400 watts
And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
Thanks for the replys, i wanted to go with a 1000w due to future upgrades crossfire ect, and i thought may aswell put in a larger psu now than later on, And hhd wise i play a hell of alot of games, make mods ect, and well my hdd runs up in no time, But then again like you said bottleneck, hmm
I always thought that the i7-920 would bottleneck the pc, cant remember where i read it but using a 5970 with a 920 would be bottlenecked because of the cpu, not sure how true that is though,,
So when we reduce the psu's efficiency from total system consumption 500*0,8 = 400 watts And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
As for the SSD setup, if you have the cash, best thing to do is SSD for OS and your main programs/games, then a 500gb-1TB hard drive for data storage and stuff that performance doesn't matter on. This is about the same price as 2 velociraptors and gives you faster performance on the SSD stuff and more storage for the data stuff.
Originally posted by dfan So when we reduce the psu's efficiency from total system consumption
500*0,8 = 400 watts
And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
Originally posted by dfan Originally posted by noquarter
Originally posted by dfan
Originally posted by noquarter Originally posted by dfan So when we reduce the psu's efficiency from total system consumption 500*0,8 = 400 watts And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall. Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
Which are furmark + prime. I don't mean to say the measurements are wrong but numbers during normal load are so much smaller. Some people have burned their vrm chips with furmark cause it gives so unrealistically heavy load for the card.
Oh, yea I agree the max wattage you see on those benchmarks are something you'll never achieve in use, but since you gotta bump up the PSU to keep it from running at max load I usually just go off the max wattage anyway.
Though looking at it again the 3d Mark 06 results are only 350W off the PSU. You could be right about that graph you posted being off the wall, I couldn't find what methodology they used for that test.
Perhaps it's best to aim for the Furmark+Prime (100% load) to be a ~90% load on your PSU.. by that criteria 500W max load under Furmark+Prime would be acceptable on even a 550W PSU, and 350W gaming usage would be a breeze for it. You'd still want at least 100W room for Crossfire though, and then some space for capacitor deterioration, so 750W is probably the best match.
But an i7 920 + 5970 is such a huge investment it doesn't seem worth it to not drop another 20 bucks to go 850w-950w and not worry about it at all.
- ATI Radeon HD 5970 (2GB) So when we reduce the psu's efficiency from total system consumption 500*0,8 = 400 watts And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Modern graphics cards demand a lot of juice, with 650W my computer wont even start (I have a 295 GTX card and multiple harddrives, 6 right now but I had more before).
750W is perfect for me and better safe than sorry. With a Corsair PSU it will only use the power it needs also, and the OP isn't going for low cost.
Better safe than sorry.
To the OP: The guy who told you to put the OS and games on a SSD and have a 500gb media drive is right, it will be faster and silent, and safer. SSDs don't crash like regular HDs (even though backing it up at times always is a good idea). The harddrives is doing a lot of noises, SSDs are both silent and doesn't generate any heat.
So when we reduce the psu's efficiency from total system consumption
500*0,8 = 400 watts
And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
Which are furmark + prime. I don't mean to say the measurements are wrong but numbers during normal load are so much smaller.
Some people have burned their vrm chips with furmark cause it gives so unrealistically heavy load for the card.
Oh, yea I agree the max wattage you see on those benchmarks are something you'll never achieve in use, but since you gotta bump up the PSU to keep it from running at max load I usually just go off the max wattage anyway.
Though looking at it again the 3d Mark 06 results are only 350W off the PSU. You could be right about that graph you posted being off the wall, I couldn't find what methodology they used for that test.
Perhaps it's best to aim for the Furmark+Prime (100% load) to be a ~90% load on your PSU.. by that criteria 500W max load under Furmark+Prime would be acceptable on even a 550W PSU, and 350W gaming usage would be a breeze for it. You'd still want at least 100W room for Crossfire though, and then some space for capacitor deterioration, so 750W is probably the best match.
But an i7 920 + 5970 is such a huge investment it doesn't seem worth it to not drop another 20 bucks to go 850w-950w and not worry about it at all.
Psus used to be much more picky about efficiency, the sweetspot was ~80% of nominal wattage. During every other load the efficiency was much worse, we are talking of amounts like even 30%. Today they work much better excluding very low loads.
700-800 is ok if you want to leave an option for CF, but supplies like 1kW are totally absurd for 99% of computers.
Don't buy ancient nvidia, newer radeon is much better choice.
Depends on what card you are going to pick, Nvidia still holds a couple of sweet spots of price/performance and overclocking potential.
It is not because that now ATI is king of the hill at the top range, that in medium range there is no competition. It is no different then when Nvidia was on top and ATI had good price/performance medium range cards.
Besides that, it took them ( ATI ) long enough to finally close the gap and take back the lead. Nvidia had the lead since the 8 series, I doubt that this time ATI will be in the lead for that long.
If you are interested in subscription or PCU numbers for MMORPG's, check out my site : http://mmodata.blogspot.be/ Favorite MMORPG's : DAoC pre ToA-NF, SWG Pre CU-NGE, EVE Online
Don't buy ancient nvidia, newer radeon is much better choice.
Depends on what card you are going to pick, Nvidia still holds a couple of sweet spots of price/performance and overclocking potential.
It is not because that now ATI is king of the hill at the top range, that in medium range there is no competition. It is no different then when Nvidia was on top and ATI had good price/performance medium range cards.
Besides that, it took them ( ATI ) long enough to finally close the gap and take back the lead. Nvidia had the lead since the 8 series, I doubt that this time ATI will be in the lead for that long.
There are many other factors to consider than just performance. Things like future driver support, power draw, dx support etc. Pretty much everything points towards ati.
Comments
Wow that's a hell of a rig. Only thing that could make it better is an SSD on top of your 1TB drive.
Also while the Cooler Master PSU would work fine, I'd look at the Corsair TX 950W or Silverstone 1000W.
Don't waste your money on the TB.
You can downgrade to the i7 950 and use the extra money to upgrade to a decent SSD. They really aren't all that bad priced anymore. If you want the TB, keep it for storage but you're bottle neck is going to be the HD. Or you can get a 500gb Western Digital Passport and have portable storage. A TB is alot of space for everything you need except movies, and a TB can't hold enough movies to make it worth it, imho.
I have a very simliar system with a 1TB 7200rpm HD.
Windows Rating:
Processor: 7.4 (i7 920 not OC)
RAM: 7.5 (6gb DDR3 1666)
Graphics: 7.3
Gaming Graphics: 7.3 (GTX 280)
Primary Hard Disk: 5.9
Uggg. I hate it.
Like Trading Card Games? Click Here.
A 600 watt psu will do, you don't need a nuclear power plant for normal computer.
I wouldn't change x58 stuff away, it's with multiple gfx cards where triple channel memory is most effective.
Wow, I actually though the 5890 had a much higher power consumption than it does. 190W for the 5890. That would have an absolute maximum power draw of 500W in an i7 920 system, so a 750W PSU would cover it easy. If he wanted to put another 5870 or 5890 into Crossfire he'd probably want a 800W-950W PSU though.
A 600W would work too but with degradation it might not quite cut it in a couple years, I like to recommend a little bit higher just to be safe so noone goes, man that guy gave horrible advice and I got screwed :P But yea 1000W are almost always uncalled for.
Wow, I actually though the 5890 had a much higher power consumption than it does. 190W for the 5890. That would have an absolute maximum power draw of 500W in an i7 920 system, so a 750W PSU would cover it easy. If he wanted to put another 5870 or 5890 into Crossfire he'd probably want a 800W-950W PSU though.
A 600W would work too but with degradation it might not quite cut it in a couple years, I like to recommend a little bit higher just to be safe so noone goes, man that guy gave horrible advice and I got screwed :P But yea 1000W are almost always uncalled for.
Here's a small example of power consumption:
Core i7 Test System Specs
- Intel Core i7 965 Extreme Edition (Overclocked @ 3.70GHz)
- x3 2GB G.Skill DDR3 PC3-12800 (CAS 9-9-9-24)
- Asus P6T Deluxe (Intel X58)
- OCZ GameXStream (700 watt)
- Seagate 500GB 7200-RPM (Serial ATA300)
- ATI Radeon HD 5970 (2GB)
So when we reduce the psu's efficiency from total system consumption
500*0,8 = 400 watts
And this is with heavily overclocked system, you can state good quality 600 watt psu is surely enough.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
Hmm you only have to reduce for the PSU's efficiency if the measurement is coming off a wall meter but given the similarity to xbitlabs' review of an i7 920 system with a GTX295 which is definitely measured from the PSU rather than from the wall I'd guess that graph is direct PSU measurement as well. So that 500W measurement would be a 625W measurement if it was done at the wall.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
Thanks for the replys, i wanted to go with a 1000w due to future upgrades crossfire ect, and i thought may aswell put in a larger psu now than later on, And hhd wise i play a hell of alot of games, make mods ect, and well my hdd runs up in no time, But then again like you said bottleneck, hmm
I always thought that the i7-920 would bottleneck the pc, cant remember where i read it but using a 5970 with a 920 would be bottlenecked because of the cpu, not sure how true that is though,,
http://www.xbitlabs.com/articles/coolers/display/system-wattage.html
As for the SSD setup, if you have the cash, best thing to do is SSD for OS and your main programs/games, then a 500gb-1TB hard drive for data storage and stuff that performance doesn't matter on. This is about the same price as 2 velociraptors and gives you faster performance on the SSD stuff and more storage for the data stuff.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
http://www.xbitlabs.com/articles/coolers/display/system-wattage.html
Which are furmark + prime. I don't mean to say the measurements are wrong but numbers during normal load are so much smaller.
Some people have burned their vrm chips with furmark cause it gives so unrealistically heavy load for the card.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
http://www.xbitlabs.com/articles/coolers/display/system-wattage.html
Which are furmark + prime. I don't mean to say the measurements are wrong but numbers during normal load are so much smaller.
Some people have burned their vrm chips with furmark cause it gives so unrealistically heavy load for the card.
Oh, yea I agree the max wattage you see on those benchmarks are something you'll never achieve in use, but since you gotta bump up the PSU to keep it from running at max load I usually just go off the max wattage anyway.
Though looking at it again the 3d Mark 06 results are only 350W off the PSU. You could be right about that graph you posted being off the wall, I couldn't find what methodology they used for that test.
Perhaps it's best to aim for the Furmark+Prime (100% load) to be a ~90% load on your PSU.. by that criteria 500W max load under Furmark+Prime would be acceptable on even a 550W PSU, and 350W gaming usage would be a breeze for it. You'd still want at least 100W room for Crossfire though, and then some space for capacitor deterioration, so 750W is probably the best match.
But an i7 920 + 5970 is such a huge investment it doesn't seem worth it to not drop another 20 bucks to go 850w-950w and not worry about it at all.
Modern graphics cards demand a lot of juice, with 650W my computer wont even start (I have a 295 GTX card and multiple harddrives, 6 right now but I had more before).
750W is perfect for me and better safe than sorry. With a Corsair PSU it will only use the power it needs also, and the OP isn't going for low cost.
Better safe than sorry.
To the OP: The guy who told you to put the OS and games on a SSD and have a 500gb media drive is right, it will be faster and silent, and safer. SSDs don't crash like regular HDs (even though backing it up at times always is a good idea). The harddrives is doing a lot of noises, SSDs are both silent and doesn't generate any heat.
Tried to look for that review but didn't find it. Afaik they have only measured gfx amps from pcie lanes. The idea of measuring total computer power straight from voltage lanes is kinda absurd imo, it would require so many multimeters. I'm quite sure the number is measured straight from socket.
Also you need to be bit skeptical when reading those numbers, most are furmark results which are much higher than those you can achieve with games and other normal load.
http://www.xbitlabs.com/articles/coolers/display/system-wattage.html
Which are furmark + prime. I don't mean to say the measurements are wrong but numbers during normal load are so much smaller.
Some people have burned their vrm chips with furmark cause it gives so unrealistically heavy load for the card.
Oh, yea I agree the max wattage you see on those benchmarks are something you'll never achieve in use, but since you gotta bump up the PSU to keep it from running at max load I usually just go off the max wattage anyway.
Though looking at it again the 3d Mark 06 results are only 350W off the PSU. You could be right about that graph you posted being off the wall, I couldn't find what methodology they used for that test.
Perhaps it's best to aim for the Furmark+Prime (100% load) to be a ~90% load on your PSU.. by that criteria 500W max load under Furmark+Prime would be acceptable on even a 550W PSU, and 350W gaming usage would be a breeze for it. You'd still want at least 100W room for Crossfire though, and then some space for capacitor deterioration, so 750W is probably the best match.
But an i7 920 + 5970 is such a huge investment it doesn't seem worth it to not drop another 20 bucks to go 850w-950w and not worry about it at all.
Psus used to be much more picky about efficiency, the sweetspot was ~80% of nominal wattage. During every other load the efficiency was much worse, we are talking of amounts like even 30%. Today they work much better excluding very low loads.
700-800 is ok if you want to leave an option for CF, but supplies like 1kW are totally absurd for 99% of computers.
This is my budget build I'm working on atm
SAMSUNG 2343BWX 23’ LCD (purchased)
Antec Twelve hundred Black steel case (purchased)
Gigabyte GA-P55-UD6 socket 1156
Intel Core i7-860 Lynnfield
CORSAIR CMPSU-850HX 850W (purchased)
Noctua NH-U12P SE2 CPU cooler (purchased)
G.SKILL Ripjaws Series 4GB DDR3 1333 (PC3 10666) (x2)
MSI TwinFrozr OC GeForce GTX 275 (may add a 2nd later)
Sennheiser PC156 headset (purchased)
LG BluRay writer (OEM)
Logitech cordless rumblePad (purchased)
LOGISYS dual UV cold cathode kit (purchased)
Antec Tri cool fans (x2) red and blue (purchased)
Western Digital Caviar Black (1TB) (may add 2nd later)
Microsoft Windows 7 Ultimate 64-bit
Don't buy ancient nvidia, newer radeon is much better choice.
Depends on what card you are going to pick, Nvidia still holds a couple of sweet spots of price/performance and overclocking potential.
It is not because that now ATI is king of the hill at the top range, that in medium range there is no competition. It is no different then when Nvidia was on top and ATI had good price/performance medium range cards.
Besides that, it took them ( ATI ) long enough to finally close the gap and take back the lead. Nvidia had the lead since the 8 series, I doubt that this time ATI will be in the lead for that long.
If you are interested in subscription or PCU numbers for MMORPG's, check out my site :
http://mmodata.blogspot.be/
Favorite MMORPG's : DAoC pre ToA-NF, SWG Pre CU-NGE, EVE Online
Depends on what card you are going to pick, Nvidia still holds a couple of sweet spots of price/performance and overclocking potential.
It is not because that now ATI is king of the hill at the top range, that in medium range there is no competition. It is no different then when Nvidia was on top and ATI had good price/performance medium range cards.
Besides that, it took them ( ATI ) long enough to finally close the gap and take back the lead. Nvidia had the lead since the 8 series, I doubt that this time ATI will be in the lead for that long.
There are many other factors to consider than just performance. Things like future driver support, power draw, dx support etc. Pretty much everything points towards ati.