• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

NVIDIA GeForce GTX 960 Launch Date Revealed

Effectively it's comparable to less skilled overclockers who overclock amd cpu's with all the power savings off, and don't know how to overclock using offset voltage modes.
The cpu never throttles but it's an inefficient way of overclocking.

Nvidia would be the amd user enabling cool and quiet,c's and APM mode on,so you get the low clock speeds and low voltages at idle and then when the cpu is needed it adjusts the p-state to the workload required, but due to APM mode on it might not hold the clocks at full potential at a constant. If you turn the APM mode off then it'll hold the full boost clocks but consumption will increase when measuring over a longer period of time.
 
Effectively it's comparable to less skilled overclockers who overclock amd cpu's with all the power savings off, and don't know how to overclock using offset voltage modes.
The cpu never throttles but it's an inefficient way of overclocking.

Nvidia would be the amd user enabling cool and quiet,c's and APM mode on,so you get the low clock speeds and low voltages at idle and then when the cpu is needed it adjusts the p-state to the workload required, but due to APM mode on it might not hold the clocks at full potential at a constant. If you turn the APM mode off then it'll hold the full boost clocks but consumption will increase when measuring over a longer period of time.

Or HPM, "High Performance Computing" I tried that, it freezes up the system eventually as it tries to pass the BPL limits. :o

Its not designed to be on in conjunction with overclocking, AMD have vivid warnings about doing that in their overclocking guides, and for good reason.

APM "Application Power Management" yes, a bit like Maxwell will monitor how much power the CPU requires and then feed it accordingly, if it requires 100% of the power it will feed the CPU that.
One could say AMD have had this technology for many years, in all likelihood it will find its way onto GPU's, and sooner rather than later.

But what remains is the fact that AMD's CPU's are designed to run at 140 Watts at a constant and will, mine does it a lot as i do a lot of rendering, ignoring the fact that i'm overclocked and have it turned off as it is not entirely stable with it on overclocked.

Maxwell is no different to that, it will get asked to run at full tilt with some things and it will oblige, as Toms Hardware demonstrated by asking it to do a simple rendering task.
 
Yeah, they are 'that' only power efficient when its not running at full steam because its able to regulate the amount of power it uses in conjunction with how much its stressed, if its at full stress it sucks just as much power as any GPU of similar performance, including the 290.

He does seem to be stuck in a cycle of defending Nvidia at all costs these days, including rubbishing arguably the bast hardware reviewer on the net ^^^^

You gave your opinion that a 970 uses as much power as a 290 when running games at similar fps, then when it looked to be incorrect you **** someone off for testing it in exactly the same way electricity companies charge us for it.

Man up and admit your wrong.

I feel sorry for the guy that asked about power consumption now as I bet he is none the wiser.

I bet the 960 variants perform similarly than the 280x variants while using less power :)
 
Last edited:
You gave your opinion that a 970 uses as much power as a 290 when running games at similar fps, then when it looked to be incorrect you **** someone off for testing it in exactly the same way electricity companies charge us for it.

Man up and admit your wrong.

I feel sorry for the guy that asked about power consumption now as I bet he is none the wiser.

No i didn't.
 
You didn't what?

1. Say the 290 uses the same power as a 970?

2. Take issues with Kaapstad's testing?

1, Say 970 can and sometimes will use as much as the 290

2, He tested 4 GPU's, overclocked, over volted, over reference power limits with the two vendors in different rigs, explain to me how that is a template for an accurate depiction of one out of the box GPU.
 
2, He tested 4 GPU's, overclocked, over volted, over reference power limits with the two vendors in different rigs, explain to me how that is a template for an accurate depiction of one out of the box GPU.


Because maths.
 
Because maths.

Alright, you do the Maths.

Product A, GPU Frequency 1230mhz, Testing methodology: Tomb Raider, Voltage: unknown: power limits: Unknown (assume +<150 Watts; +50%), CPU Voltage: unknown. Average power draw 1700 Watts, low power draw: Unknown, high power draw: unknown
Separate the 4 GPU's from the differing Rigs, divide that by 4.

Product B, GPU Frequency 1450mhz, Testing methodology: Tomb Raider, Voltage: unknown: power limits: Unknown (assume +6%; Wattage unknown ), CPU Voltage: unknown, Average power draw 1100 Watts, low power draw: Unknown, high power draw: unknown
Separate the 4 GPU's from the differing Rigs, divide that by 4.

Now give me the low, average and high power draw of one of those 4 GPU's when not overclocked, over volted and with over reference power limits set.

go...
 
Last edited:
1, Say 970 can and sometimes will use as much as the 290

2, He tested 4 GPU's, overclocked, over volted, over reference power limits with the two vendors in different rigs, explain to me how that is a template for an accurate depiction of one out of the box GPU.

1. At similar performance levels you said! And the 970 uses less.

2. His amd rig uses less power Than the rig with the 980's if both had no cards in, TheN Add the cards and the amd one uses even more power while being slower than the nvidia one.

So not sure what you saying.
 
It is actually more scientific to use 4 GPUs as this reduces the influence of the rest of the system on what the GPUs are using.

If I had used 1 GPU then out of the total power used the rest of the system would have used a far higher percentage masking what the GPU was doing.:)

@humbug when I pay my electricity bill it is based on the metering the elec board uses not an oscilloscope.:D

I do hope Tom's did the article to help users make an informed choice when it comes to the elec bill.:)
 
Oh well I tried. :rolleyes:



Lets just get one thing straight.

All these numbers mean nothing, absolutely nothing. Why I hear you ask.

Well for a start everybody will play each game differently whereas humbug and andy might be of similar age to me I can assure you that their reflexes are way better than mine, so for that same 1 hours gameplay I will spend a lot more of it dead than they will so the GPU wont be working as hard rendering what is going on all the time.:D

You really want to know how much energy you Computer uses ?
go take an reading off of your electricity meter, then sit there doing whatever you normally do (but not on the computer) for a set period (3 hours for example) go take another reading, now go do what you normally do but play games for the same period, take a reading.
That is how much your computer uses and it is really the only number you have to worry about as that is the numbers you will be billed for. Good luck arguing with your electricity supplier that the measuring system they use isn't accurate enough.

Anyone can pull numbers together to say what they want and there will always be someone to disagree with them.

Also 84% of stats on the internet are made up if you didn't know. ;)


Now can we get back to talking about the 960 and don't even try to say that the last few pages has been about the 960 because that is billyhooks, you guys have just been arguing because you don't know how to stop.:mad:
 
It is actually more scientific to use 4 GPUs as this reduces the influence of the rest of the system on what the GPUs are using.

If I had used 1 GPU then out of the total power used the rest of the system would have used a far higher percentage masking what the GPU was doing.:)

@Humbug when I pay my electricity bill it is based on the metering the elec board uses not an oscilloscope.:D

I do hope Tom's did the article to help users make an informed choice when it comes to the elec bill.:)

When you set AMD's power limits to +50 the GPU will potentially pull anything upto 150 Watts more power if it needs it or not.

As far as i'm aware Nvidia still only allow +6%, assuming a BIOS power limit of 300 Watts that +18 Watts
that has a very real potential for the AMD GPU's pulling far more power than they normally would.
In my CPU analogy Its the same as setting HPC and setting higher power limits to allow it to keep pulling more and more power regardless of what it needs.
While the Nvidia GPU is much more limited in its power draw, by some 130 Watts. A completely unnecessary thing to do, perhaps this is why my GPU runs so cool and quiet given that i do not ever set anywhere near that higher power limit

By using 4 GPU's you introduce an element of CPU bottlenecking, that will allow the Nvidia GPU's to run in a lower power state given its APM stile firmware, AMD GPU's have no such firware and will power themselves at the high end continuously.

i'm not saying any or all of that is whats going on, but it is some of the many variables that are included.
 
Last edited:
Wait, so a CPU bottleneck on the AMD cards wont stop them running at 100% load? Damn, so inefficient. They should probably have some form of dynamic load thing. Killing all those polar bears. AMD are a bad company.
 
When you set AMD's power limits to +50 the GPU will potentially pull anything upto 150 Watts more power if it needs it or not.

As far as i'm aware Nvidia still only allow +6%, assuming a BIOS power limit of 300 Watts that +18 Watts
that has a very real potential for the AMD GPU's pulling far more power than they normally would.
In my CPU analogy Its the same as setting HPC and setting higher power limits to allow it to keep pulling more and more power regardless of what it needs.
While the Nvidia GPU is much more limited in its power draw, by some 120 Watts.

By using 4 GPU's you introduce an element of CPU bottlenecking, that will allow the Nvidia GPU run run in a lower power state given its APM stile firmware, AMD GPU have no such firware and will power themselves at the high end continuously.

i'm not saying any or all of that is whats going on, but it is some of the many variables that are included.

Give up Humbug you are busted on this.:D

I also ran the 980s @2160k as well because I knew you were going to mention bottlenecking and the end result was there is not much of an increase in power usage. Even overclocked @4K the 980s use less power than the 290Xs @1080p @stock running lower settings.:D

You may also want to look at the settings I used for my 980s again, I did post them several pages back for your benefit but you obviously did not read them.:)

0vL1Iab.jpg
 
Give up Humbug you are busted on this.:D

I also ran the 980s @2160k as well because I knew you were going to mention bottlenecking and the end result was there is not much of an increase in power usage. Even overclocked @4K the 980s use less power than the 290Xs @1080p @stock running lower settings.:D

You may also want to look at the settings I used for my 980s again, I did post them several pages back for your benefit but you obviously did not read them.:)

0vL1Iab.jpg

Kaap, what is the low, Average and high power consumption of one GPU running at 1440P, without overclock, without over volting, without adding 150 Watts ontop of the board power line.
 
Last edited:
He's already said, the 290s @stock use more than overclocked/overvolted and running@4k 980s. The 290s were @1080 so potential CPU bottleneck saving power.

I don't know why you're still clinging onto this.

He's still using 4, still with overclocks, still with over volting, still with added 150 Watt ontop of the AMD GPU's BPL. and still only know what his averages are.
 
Back
Top Bottom