Power, Temperature, & Noise

As always, we wrap up our look at a new video card with a look at the physical performance attributes: power consumption, temperatures, and noise. With new process nodes being the lifeblood of the GPU industry, each new process gives GPU manufacturers a chance to move their product along a continuum; do they take advantage of a power consumption reduction for the same performance level, a performance increase for the same power consumption, or something in between? In AMD’s case they’ve chosen to try to maximize performance within Cayman’s power budget, which means power, temperature, and noise should be similar to what we’ve seen with 6970, cooler improvements not withstanding.

Before we get into the charts, it’s worth noting right now that we don’t have a good idea of what 7970’s operational voltage is, as we don’t have any tools that can read 7970’s VRMs. We believe it’s close to Cayman’s, but this is something that will require confirmation in the future.

Idle Power Consumption

Starting as always with idle power, we can immediately see the benefits of the work AMD has put into idle power usage with Southern Islands. AMD has been working hard to reduce idle power consumption since the 4870 and at this point they’ve finally beaten even the 3870, which was a relatively small GPU using GDDR3 RAM. AMD’s official typical idle TDP here is 15W, and we have little doubt they’re going to continue to whittle that number down on future generations.

Long Idle Power Consumption

While we’re looking at idle power, we also wanted to take a look at AMD’s “long idle” scenario, where a blanked display or being slave GPU in a multi-GPU setup gives a GPU the opportunity to go into a deeper sleep state as it’s not needed for any work. Through ZeroCore Power AMD is able to shut off virtually every aspect of Tahiti when in a long idle state, allowing AMD to reduce the power consumption of 7970 to under 3W. Meanwhile for every other card there’s a very slight gain to be had in long idle because the GPU can power down all of its display transmitters, but it’s not nearly as effective as shutting down the entire GPU, which is why 7970 has a 10W advantage at the wall versus the next closest GPU.

As we said earlier in our look at ZeroCore Power technology, while the numbers are impressive enough, for the desktop the real use will be in multi-GPU systems as slave GPUs can be put in a ZCP state even while the master GPU is awake and working, significantly reducing idle power and noise in multi-GPU systems.

Load Power Consumption - Metro 2033

Moving on to our load power tests, based on our testing we have swapped out Crysis for Metro 2033 in this iteration of our benchmark suite as Metro 2033 was proving to be a bit more strenuous on our GPUs and is the only game to ever trigger PowerTune on the 6970.

Looking at Metro it’s clear that the smaller fabrication process for 6970 has allowed AMD to achieve a lower power consumption level relative to their gaming performance – the 7970 leads the GTX 580 by 20-30% in performance while consuming 34W less at the wall. At the same time this test calls into question AMD’s power targets. We’re clearly drawing more power than the 6970 – 21W more to be precise – and while this could be explained by CPU power consumption I’m not convinced that’s the case.

Load Power Consumption - OCCT

As for our pathological power consumption test we’ve moved on to OCCT, which at this point in time escapes NVIDIA’s throttling mechanism and in spite of the presence of PowerTune still manages to be a hardware workload on AMD’s GPUs than FurMark. OCCT backs up our earlier data from Metro that 7970 draws more power than 6970 even with the similar power targets. We’re drawing 50W more at the wall, which on paper at least would put 7970’s power consumption closer to 280W if AMD’s original 250W target for 6970 is still correct. If all of this is correct, then it calls into doubt AMD’s published power targets, as it looks like AMD has made a power/performance tradeoff with 7970 by trading slightly higher power consumption for what would be higher performance.

Idle GPU Temperature

Up next is our look at idle temperatures. All of the last generation GPUs were highly competitive here and as a result it’s a tough crowd as there’s relatively little room to further reduce a GPU’s temperature. Nevertheless the 7970 manages to squeak by the competition, with a recorded idle temperature of 34C, 1C lower than the GTX 580 and 2C lower than the 6970.

Load GPU Temperature - Metro 2033

Moving on to temperatures under Metro, the results are quite surprising: the 7970 does really, really good here given its power consumption. At 74C the only cooler video cards are the 6850 and 5850, both of which are lower tier, lower power versions of higher end parts. Part of this can be traced back to AMD’s cooling optimizations such as moving back to a full slot exhaust vent, but I don’t believe that’s a full explanation. We’ll get back to this when we’re looking at noise.

Load GPU Temperature - OCCT

While the overall numbers are higher in OCCT, the relative rankings are unchanged from Metro. The only cooler cards in this lineup are the 6950 and 5850. Here the 7970 hits 79C, while the 6970 is slightly behind at 80C, and the GTX 500 series is in the upper 80s.

Idle Noise Levels

Last we have our look at noise, starting with idle noise. Much like our idle temperatures most single-GPU video cards tend to cluster together within 1dB, but the 7970 has managed to silence most of the rest of the pack, with a recorded noise level of 40.2dB. More impressive perhaps is the 7970 when in its Zero Core Power state. Because the fan turns off, the card is silent; 37.9dB(A) is the noise floor of our testbed.

Load Noise Levels - Metro 2033

Moving on to our load noise measurements the full picture of the 7970’s cooling performance finally starts coming together. We’ve seen that the 7970 is a cool card, but is it a quiet card? The numbers say no. The 7970 is 3dB louder than the GTX 580 and 2dB louder than the 6970 even with its physical cooling improvements. Based on this combined with our temperature data it’s clear that AMD has targeted aggressive cooling over silence, whereas NVIDIA traditionally targets silence over aggressive cooling. When it comes to our testbed AMD has probably overdone it – we could easily exchange 10C for less noise – but with a wide range of computers in the world it’s hard to tell from a single testbed whether the cooling here is genuinely too aggressive for everyone, or if it’s the opposite. What isn’t in question is that the 7970 is going to be moderately loud, which coming from AMD is a surprising development.

Load Noise Levels - OCCT

Once again with OCCT the numbers are higher, but the facts are largely the same. The 7970 is now quieter than the GTX 580 due to PowerTune, but it’s now 4dB louder than the 6970. Unlike Metro 57.3dB of noise is in the middle of the pack, but it stands to reason that AMD could have been a bit less aggressive on their fan speeds and made a quieter card as a result.

Theoreticals & Tessellation Final Words
Comments Locked

292 Comments

View All Comments

  • CeriseCogburn - Thursday, March 8, 2012 - link

    Interesting, amd finally copied nvidia...
    " This problem forms the basis of this benchmark, and the NQueen test proves once more that AMD's Radeon HD 7970 tremendously benefits from leaving behind the VLIW architecture in complex workloads. Both the HD 7970 and the GTX 580 are nearly twice as fast as the older Radeons. "

    When we show diversity we should also show that amd radeon has been massively crippled for a long time except when "simpleton" was the key to speed. "Superior architecture" actually means "simple and stupid" - hence "fast" at repeating simpleton nothings, but unable to handle "complex tasks".
    LOL - the dumb gpu by amd has finally "evolved".
  • chizow - Thursday, December 22, 2011 - link

    ....unfortunately its going to be pitted against Kepler for the long haul.

    There's a lot to like about Southern Islands but I think its going to end up a very similar situation as Evergreen vs. Fermi, where Evergreen released sooner and took the early lead, but Fermi ultimately won the generation. I expect similar with Tahiti holding the lead for the next 3-6 months until Kepler arrives, but Kepler and its refresh parts winning this 28nm generation once they hit the streets.

    Overall the performance and changes AMD made with Tahiti look great compared to Northern Islands, but compared to Fermi parts, its just far less impressive. If you already owned an AMD NI or Evergreen part, there'd be a lot of reason to upgrade, but if you own a Fermi generation Nvidia card there's just far less reason to, especially at the asking price.

    I do like how AMD opened up the graphics pipeline with Tahiti though, 384-bit bus, 3GB framebuffer, although I wonder if holding steady with ROPs hurts them compared to Kepler. It would've also been interesting to see how the 3GB GTX 580 compared at 2560 since the 1.5GB model tended to struggle even against 2GB NI parts at that resolution.
  • ravisurdhar - Thursday, December 22, 2011 - link

    My thoughts exactly. Can't wait to see what Kepler can do.

    Also...4+B transistors? mind=blown. I remember when we were ogling over 1B. Moore's law is crazy.... :D
  • johnpombrio - Wednesday, December 28, 2011 - link

    Exactly. If you look at all the changes that AMD did on the card, I would have expected better results: the power consumption decrease with the Radeon 7970 is mainly due to the die shrink to 28nm. NVidia is planning on a die shrink of their existing Fermi architecture before Kepler is released:

    http://news.softpedia.com/news/Nvidia-Kepler-Is-On...

    Another effect of the die shrink is that clock speed usually increases as there is less heat created at the lower voltage needed with a smaller transistor.

    The third change that is not revolutionary is the bump of AMD's 7970's memory bus from 384 bits (matching the 580) from the 6970's 256 bits along with 3GB DDR5 memory vs the GTX580's 1.5GB and the 6970's 2GB.

    The final non revolutionary change is bumping the number of stream processors by 33% from 1,536 to 2,048.

    Again, breaking out my calculator, the 35% bump in the number of stream processors ALONE causes the increase in the change in the benchmark differences between the 7970 and the 6970.

    The higher benchmark, however, does not show ANY OTHER large speed bumps that SHOULD HAVE OCCURED due to the increase in the memory bus size, the higher amount of memory, compute performance, texture fill rate, or finally the NEW ARCHITECTURE.

    If I add up all the increases in the technology, I would have expected benchmarks in excess of 50-60% over the previous generation. Perhaps I am naive in how much to expect but, hell, a doubling of transistor count should have produced a lot more than a 35% increase. Add the new architecture, smaller die size, and more memory and I am underwhelmed.
  • CeriseCogburn - Thursday, March 8, 2012 - link

    Well, we can wait for their 50%+ driver increase package+ hotfixes - because after reading that it appears they are missing the boat in drivers by a wide margin.
    Hopefully a few months after Kepler blows them away, and the amd fans finally allow themselves to complain to the proper authorities and not blame it on Nvida, they will finally come through with a "fix" like they did when the amd (lead site review mastas) fans FINALLY complained about crossfire scaling....
  • KaarlisK - Thursday, December 22, 2011 - link

    What is the power consumption with multiple monitors? Previously, you could not downclock GDDR5, so the resulting consumption was horrible.
  • Ryan Smith - Thursday, December 22, 2011 - link

    "On that note, for anyone who is curious about idle clockspeeds and power consumption with multiple monitors, it has not changed relative to the 6970. When using a TMDS-type monitor along with any other monitor, AMD has to raise their idle clockspeeds from 350MHz core and 600Mhz memory to 350MHz core and the full 5.5GHz speed for memory, with the power penalty for that being around 30W. Matched timing monitors used exclusively over DisplayPort will continue to be the only way to be able to use multiple monitors without incurring an idle penalty."
  • KaarlisK - Thursday, December 22, 2011 - link

    Thank you for actually replying :)
    I am so sorry for having missed this.
  • ltcommanderdata - Thursday, December 22, 2011 - link

    Great review.

    Here's hoping that AMD will implement 64-bit FP support across the whole GCN family and not just the top-end model. Seeing AMD's mobile GPUs don't use the highest-end chip, settling for the 2nd highest and lower, there hasn't been 64-bit FP support in AMD mobile GPUs since the Mobility HD4800 series. I'm interested in this because I can then dabble in some 64-bit GPGPU programming on the go. It also has implications for Apple since their iMacs stick to mobile GPUs, so would otherwise be stuck without 64-bit FP support which presumably could be useful for some of their professional apps.

    In regards to hardware accelerated Megatexture, is it directly applicable to id Tech 5's OpenGL 3.2 solution? ie. Will id Tech 5 games see an immediate speed-up with no recoding needed? Or does Partially Resident Texture support require a custom AMD specific OpenGL extension? If it's the later, I can't see it going anywhere unless nVidia agrees to make it a multivendor EXT extension.
  • Ryan Smith - Thursday, December 22, 2011 - link

    Games will need to be specifically coded for PRT; it won't benefit any current games. And you are correct in that it will require and AMD OpenGL extension to use (it won't be accessible from D3D at this time).

Log in

Don't have an account? Sign up now