Overclock.net › Forums › Graphics Cards › Graphics Cards - General › 7950 or gtx 670?
New Posts  All Forums:Forum Nav:

7950 or gtx 670? - Page 11

post #101 of 237
Every time I see a thread discussing the 7950, I can only think of this:

451

That being said, I sadly didn't realize that the 670 was available yet. redface.gif I'd probably go with the 670 because of the slightly lower power draw. I would be surprised if you noticed much of a performance difference between the two.
    
CPUMotherboardGraphicsRAM
i5-2500K Biostar TP67B+ XFX HD5750 1GB 2x4GB DDR3 Corsair 1600 
Hard DriveOSMonitorPower
60GB OCZ SSD, 2x160GB HDD RAID0, 500GB+500GB+1.5TB Windows 7 Ultimate 64-bit Samsung SyncMaster 930B Antec SmartPower 450w 
Case
Antec 900 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5-2500K Biostar TP67B+ XFX HD5750 1GB 2x4GB DDR3 Corsair 1600 
Hard DriveOSMonitorPower
60GB OCZ SSD, 2x160GB HDD RAID0, 500GB+500GB+1.5TB Windows 7 Ultimate 64-bit Samsung SyncMaster 930B Antec SmartPower 450w 
Case
Antec 900 
  hide details  
Reply
post #102 of 237
Quote:
Originally Posted by guyladouche View Post

That being said, I sadly didn't realize that the 670 was available yet. redface.gif I'd probably go with the 670 because of the slightly lower power draw. I would be surprised if you noticed much of a performance difference between the two.

It's not slight, it's a massive difference at a full overclock. It's at least 28% less to be more precise which is about $30 less per year at 5 hours of load per day at a full overclock.
Edited by SeanPoe - 5/14/12 at 4:40pm
post #103 of 237
Quote:
Originally Posted by SeanPoe View Post

This is definitely correct, i even made the foolish mistake of doing it the other way here when i calculated it (calculating default core speed vs oc core speed (without boost, assuming this to be a nearly negligible constant) as the percent of overclock, rather than boosted core vs OC boosted core).
We don't have enough data to calculate it properly because no one is listing their stock boosted core speeds. So all we know is the OC boosted core speeds. But from what i've seen, the boost varies between +140-180 which means you can calculate the scaling (like i did in the link above in this post) by just assuming the boost is +~150 for both the default and OC core values. It's not 100% accurate but it should be accurate within 3-6% is my guess given the range on the boost.

I concur thumb.gif

However as I'm sure you realize, doing it by your method will always under-report the actual scaling % ... but 3-6% sounds reasonable as a range ... the higher the observed boost, the greater the discrepancy will be ... between using the correct formula vs your formula wink.gif
Quote:
Originally Posted by SeanPoe View Post

It's not slight, it's a massive difference at a full overclock. It's at least 28% less to be more precise which is about $180 less per year at 5 hours of load per day at a full overclock.

Such a strident and precise assertion of facts like 'at least 28%' should surely arrived backed by proof.gif of some kind, don't you think? Particularly when you're refuting someone directly ...

Furthermore, electricity costs are highly regional in nature. Thus, there's no way to come up w/an absolute dollar amount. Perhaps you meant to say 'where I live'?
Edited by brettjv - 5/14/12 at 3:34pm
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #104 of 237
Quote:
Originally Posted by brettjv View Post

I concur thumb.gif
However as I'm sure you realize, doing it by your method will always under-report the actual scaling % ... but 3-6% sounds reasonable as a range ... the higher the observed boost, the greater the discrepancy will be ... between using the correct formula vs your formula wink.gif
Such a strident and precise assertion of facts like 'at least 28%' should surely arrived backed by proof.gif of some kind, don't you think? Particularly when you're refuting someone directly ...
Furthermore, electricity costs are highly regional in nature. Thus, there's no way to come up w/an absolute dollar amount. Perhaps you meant to say 'where I live'?

I saw his formula and he bases his assumption on you running the card full bore 5 hours every day for an entire year which for me and I'd guess most is an unrealistic scenario. It's also based off power consumption in one game and as we know with the 680/7970 there are actually odd examples out there where the 6** series sucks more juice then the 7k. His at least statement should read at most.

Anyway point is flawed method is flawed.
    
CPUMotherboardGraphicsRAM
6700k (EK Supremacy Evo) *4.7 @ 1.312 Asus Maximus VIII gene GTX 1080 FE (EK nickel/plexi)* 2075/5401 16gb Corsair 3200  
Hard DriveHard DriveHard DriveCooling
Samsung 950pro M.2 256gb Samsung 850 Evo - 512 - Steam Samsung 850 Evo - 512 - Other Apps EK Res 
CoolingCoolingCoolingOS
EK pump top XSPC EX 360 XSPC EX 240 10 
KeyboardPowerCaseMouse
Ducky Shine 3 TKL  EVGA 650 Fractal Arc Mini 2 G303 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
6700k (EK Supremacy Evo) *4.7 @ 1.312 Asus Maximus VIII gene GTX 1080 FE (EK nickel/plexi)* 2075/5401 16gb Corsair 3200  
Hard DriveHard DriveHard DriveCooling
Samsung 950pro M.2 256gb Samsung 850 Evo - 512 - Steam Samsung 850 Evo - 512 - Other Apps EK Res 
CoolingCoolingCoolingOS
EK pump top XSPC EX 360 XSPC EX 240 10 
KeyboardPowerCaseMouse
Ducky Shine 3 TKL  EVGA 650 Fractal Arc Mini 2 G303 
  hide details  
Reply
post #105 of 237
Quote:
Originally Posted by SeanPoe View Post

It's not slight, it's a massive difference at a full overclock. It's at least 28% less to be more precise which is about $180 less per year at 5 hours of load per day at a full overclock.

(A) I wasn't talking about overclocks, (B) I doubt your numbers (if you ran one card consuming 200w all day everyday for a year and your power rate was $0.15/kw-hr, it would cost ~$260 to run that card for a whole year. If you ran a card consuming 300w all day everyday for a year it would cost $394. Even running those two different scenarios ALL DAY EVERY DAY FOR A YEAR, you still only arrive at ~$130 difference between them. If you ran them for significantly less time like 5 hours per day, you would notice hardly anything. Running a 300w card for 5 hours a day every day and paying ~ $0.15 kw-hr, you'd only pay $82 for the entire year to power the card. I don't see how it's physically possible to have a $180 difference between powering two different cards over the course of a year if you only use them for 5 hours a day).

and (C) even if there is a numbers difference, I reiterate my literal statement below:
Quote:
I would be surprised if you noticed much of a performance difference between the two.

Edited by guyladouche - 5/14/12 at 3:57pm
    
CPUMotherboardGraphicsRAM
i5-2500K Biostar TP67B+ XFX HD5750 1GB 2x4GB DDR3 Corsair 1600 
Hard DriveOSMonitorPower
60GB OCZ SSD, 2x160GB HDD RAID0, 500GB+500GB+1.5TB Windows 7 Ultimate 64-bit Samsung SyncMaster 930B Antec SmartPower 450w 
Case
Antec 900 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5-2500K Biostar TP67B+ XFX HD5750 1GB 2x4GB DDR3 Corsair 1600 
Hard DriveOSMonitorPower
60GB OCZ SSD, 2x160GB HDD RAID0, 500GB+500GB+1.5TB Windows 7 Ultimate 64-bit Samsung SyncMaster 930B Antec SmartPower 450w 
Case
Antec 900 
  hide details  
Reply
post #106 of 237
I laughed my arse off when I saw that the 670 is $520 minimum at PCCG and Umart and then the killer was you can still get a ref 7970 for $509 at both places.
post #107 of 237
Quote:
Originally Posted by brettjv View Post

Such a strident and precise assertion of facts like 'at least 28%' should surely arrived backed by proof.gif of some kind, don't you think? Particularly when you're refuting someone directly ...
Furthermore, electricity costs are highly regional in nature. Thus, there's no way to come up w/an absolute dollar amount. Perhaps you meant to say 'where I live'?

The proof is many pages back, i assumed anyone this far in this thread would have (or should) have read the majority of the posts tongue.gif. So i didn't think i needed to specifically re-say something that i already said in this thread.

Here (and i explain exactly how i got to that number) is an excerpt:
Quote:
Hardocp review analyses:
670 page
7950 page

Power consumption when overclocked under load: (just the gpu, not total system)
670: 265w
7950: 369w

That's a difference of 28% in favor of the 670 or 104w underload. This comes out to a $0.0156 difference at .15cent/kw-hr per hour under load. So it costs $0.078 more per day if used for 5 hours, or $2.34 per month, or $28.47 per year. Keep this in mind when comparing them by cost. The 7950 is 28% more expensive to run than a 670 so even if the 7950 was $371 compared to the 670 at $400, the 670 is cheaper after one year of use. But the thing is, the 7950 is currently the same cost as the 670, so it's slightly more expensive. This is why a blind price vs price comparison is mathematically incorrect, you must take into account the operation cost too. If we were to be fair to the 670 in an equal price comparison (price vs performance standpoint), then we should be comparing it to a 7870 which costs about $350.

I point out the difference between the two (28%), i point out the watt difference (104w underload) and in bold i list the exact price of power i'm using to calculate my numbers so anyone can easily calculate the saving they'd see based on their usage and their local power cost. This also doesn't take into account the 670's dynamic boost overclock which saves additional power that's not being calculated here, so worst case scenario given this data, the difference is 28% but it might be even more.

Edit: be sure you change the page to power consumption for each of those links ^_^
Edited by SeanPoe - 5/14/12 at 4:45pm
post #108 of 237
Quote:
Originally Posted by SeanPoe View Post

The proof is many pages back, i assumed anyone this far in this thread would have (or should) have read the majority of the posts tongue.gif. So i didn't think i needed to specifically re-say something that i already said in this thread.
Here (and i explain exactly how i got to that number) and excerpt:
-- snip --
I point out the difference between the two (28%), i point out the watt difference (104w underload) so anyone can easily calculate the saving they'd see based on their usage. This also doesn't take into account the 670's dynamic boost overclock which saves an additional power that's not being calculated here, so worst case scenario given this data, the difference is 28% but it might be even more.

Yes, okay, but ... doesn't it strike you as unlikely that your 28% calculation could be accurate, given it's so WILDLY different from TPU's analysis of the situation, when they used like 15 different games x 4 different resolutions?

Granted these are reference cards, but surely you're not suggesting that OC'ing alone could possibly account for an 14X LARGER difference in Perf per Watt (2% vs 28%)?

700

What do you think the odds are, statistically-speaking, of you 'getting it right' in terms of the 'population', based on a sample size of 1 benchmark, vs the chances that TPU's is more correct ... when they used well over a dozen?

I hate to tell you ... but it's nigh on impossible wink.gif
Edited by brettjv - 5/14/12 at 4:29pm
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
xeon X5675 6-core @ 4.1ghz (1.29v, 20x205 +ht ) rampage iii extreme msi rx470 gaming X (the $159 budget king) 3 x 2gb corsair xms3 pc12800 (9-9-9-24-1T@1600MHz) 
Hard DriveOptical DriveCoolingOS
hynix 250gb ssd (boot), 2tb deskstar (apps),1tb... plextor px-712sa - still the best optical drive... corsair h8o v2 aio W10 home 
MonitorPowerCaseAudio
asus vw266h 25.5" (1920x1200) abs sl (enermax revolution) * single 70A rail 850w silverstone rv-03 XFi Titanium 
  hide details  
Reply
post #109 of 237
Quote:
Originally Posted by brettjv View Post

Yes, okay, but ... doesn't it strike you as unlikely that your 28% calculation could be accurate, given it's so WILDLY different from TPU's analysis of the situation, when they used like 15 different games? Granted these are reference cards, but surely you're not suggesting that OC'ing alone could possibly account for an 18X LARGER difference in Perf per Watt (2% vs 28%)?

That's performance per watt, i was calculating just the power draw difference in percent. By this i just mean, the OC'ed 670 uses 265w, the OC'ed 7950 uses 369w when both are at full load, that's a 28% reduction in power consumption.

Edit: Using that same source as i linked above you can see that overclocking the 7950 increased it's total watts from 235 to 369 (134 increase) vs the 670 which uses 252 at stock and 265 at overclock (13w increase). Overclocking the 670 apparently doesn't increase wattage draw by much over stock given how the volts barely increase when overclocked compared to the 7950 which needs a massive volt increase.
Edited by SeanPoe - 5/14/12 at 4:22pm
post #110 of 237
Quote:
Originally Posted by SeanPoe View Post

The proof is many pages back, i assumed anyone this far in this thread would have (or should) have read the majority of the posts tongue.gif. So i didn't think i needed to specifically re-say something that i already said in this thread.
Here (and i explain exactly how i got to that number) is an excerpt:
Quote:
Hardocp review analyses:
670 page
7950 page
Power consumption when overclocked under load: (just the gpu, not total system)
670: 265w
7950: 369w
That's a difference of 28% in favor of the 670 or 104w underload. This comes out to a 10cent (.01$) difference at .15cent/kw-hr per hour under load. So it costs $0.5 more per day if used for 5 hours, or $14.5 per month, or $182.7 per year. Keep this in mind when comparing them by cost. The 7950 is 28% more expensive to run than a 670 so even if the 7950 was $217 compared to the 670 at $400, the 670 is cheaper after one year of use. But the thing is, the 7950 is currently the same cost as the 670, so it's VASTLY more expensive. This is why a blind price vs price comparison is mathematically incorrect, you must take into account the operation cost too. If we were to be fair to the 670 in an equal price comparison (price vs performance standpoint), then we should be comparing it to a 7850 or 7870 which costs about the same in the end.
I point out the difference between the two (28%), i point out the watt difference (104w underload) and in bold i list the exact price of power i'm using to calculate my numbers so anyone can easily calculate the saving they'd see based on their usage and their local power cost. This also doesn't take into account the 670's dynamic boost overclock which saves additional power that's not being calculated here, so worst case scenario given this data, the difference is 28% but it might be even more.
Edit: be sure you change the page to power consumption for each of those links ^_^
Look dude, your math is wrong from when you said there was a $180 difference in operating the two cards over the course of a year. Here's how:

265w = 0.265kw. (0.265kw) * ($0.15 kw-hr) * (5hr/day) * (365 days) = $72.54 To power the card for a whole year.

369w = 0.369kw. (0.369kw ) * ($0.15 kw-hr) * (5hr/day) * (365 days) = $108.40 To power the card for a whole year.

That is NOT $180 difference per year, that's only $35 difference per year. % difference aside, there's really hardly any monetary difference over the course of a power bill for an entire year for a household.
    
CPUMotherboardGraphicsRAM
i5-2500K Biostar TP67B+ XFX HD5750 1GB 2x4GB DDR3 Corsair 1600 
Hard DriveOSMonitorPower
60GB OCZ SSD, 2x160GB HDD RAID0, 500GB+500GB+1.5TB Windows 7 Ultimate 64-bit Samsung SyncMaster 930B Antec SmartPower 450w 
Case
Antec 900 
  hide details  
Reply
    
CPUMotherboardGraphicsRAM
i5-2500K Biostar TP67B+ XFX HD5750 1GB 2x4GB DDR3 Corsair 1600 
Hard DriveOSMonitorPower
60GB OCZ SSD, 2x160GB HDD RAID0, 500GB+500GB+1.5TB Windows 7 Ultimate 64-bit Samsung SyncMaster 930B Antec SmartPower 450w 
Case
Antec 900 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Graphics Cards - General
Overclock.net › Forums › Graphics Cards › Graphics Cards - General › 7950 or gtx 670?