2Media Player Classic - Home Cinema LogoComparing madvr GPU performance

My current HTPC has an nVidia GeForce GTX 1050Ti 4Gb with which I am able to upscale all content to 1080p with fairly high quality settings, and am also (just) able to playback 8k 60fps content downscaled to 1080p. However, I have long wondered what the benefits of a more powerful card might be on the ability to use higher settings. Sadly, no-one seems to provide any benchmarks for madvr, so how can cards be compared?


After much searching, it seems that the strongest indicator of madvr performance is the shader performance of a card. This is effectively measured in FLOPS. There are some limitations to this as a doubling of shader performance doesn’t necesarily mean a halving of render time so the relationship isn’t necesarily linear, but once other factors (e.g. memory quantity and speed) are equal, the shader performance will be a good guide to the performance of a card.

It is highly recommended to get a GPU with at least 4Gb of VRAM for madvr. Assuming a new card is being obtained, the memory is likely to be at least GDDR5, and may well be GDDR6. In these circumstances the shader performance is a useful guide.

I also used a few specific pieces of information to compare different benchmarks to see which type appeared to be the best fit. One piece of info I picked up is that my 1050Ti is faster than the Radeon RX 570 in madvr which is not reflected in lots of GPU compute benchmarks which suggests these are not a good type of benchmark to use.

Another direct comparison suggests that the 1060Ti is faster in madvr than the RX580 so this should also be reflected in any benchmarks used, which rules out most of them! In the same thread there is a suggestion that the Radeon RX 5700 XT is similar in performance to a GeForce RTX 2070 Super

Some other facts from the doom9 forums

  • The AMD RX480 is slightly faster than the GTX970
  • The GTX1070 is around 30% faster than the RX480

As I wasn’t able to find any benchmark that matched these data points, instead I’ve started by looking at all the cards available in the £100 – £200 bracket which are…

  • GTX 1660 £195
  • RX590 £190
  • RX5500 XT £160
  • GTX 1650 Super £150
  • RX580 £145
  • GTX 1050Ti £130

Is there a surprising winner, or does the price predict the performance?

The info I can find suggests that the GTX 1650 (non super) is around 25% faster than the GTX 1050Ti in madvr, and the 1650 Super is faster again (more cuda cores and a higher clock speed). It also appears the the RX5700 XT is a little faster than the GTX 1080 and that the RX5700 (non XT) is similar to a GTX 1070.

After lots of searching to compare these cards I found a very useful post on reddit which ranks the nVidia cards approximately. These rankings (unsurprisingly) do follow the price of the nVidia cards.

  • GTX 1050 / 1050 Ti
  • GTX 1650
  • GTX 1650 Super
  • GTX 1060
  • GTX 1070 / 1660 / 1660 Super / 1660 Ti
  • GTX 1080 / RX 2060
  • GTX 1080Ti / RTX 2070

The better cards (towards the bottom) tend to be beyond the price range. I can see a GTX 1070 on ebay for around £230 which is more than a new GTX 1660 so second hand options don’t appear to help. There are also 6Gb GTX 1060’s for around £170 on ebay (3Gb isn’t enough, 4Gb is a minimum really so be careful with these) – but are they really £20 better than a new GTX 1650 Super?

It sounds like the AMD Polaris generation cards weren’t good performers in madvr. This has apparently changed with Navi, however there have been driver issues with Navi so most people still recommend nVidia cards for madvr. On top of this it is rumoured that future versions of madvr may make use of the nVidia turing architecture, but these rumours are at least a year old with nothing to show so far. If they ever come true it may be that AMD’s “ray-tracing” compatible cards may be just as well suited too.

I have a suspicion that the RX 5500 XT may be a good value deal as it may offer GTX 1660 like performance in madvr, for £35 less, but I can’t find concrete evidence or a direct comparison.

What do you think? Drop us a comment below! If you would like to subscribe please use the subscribe link on the menu at the top right. You can also share this with your friends by using the social links below. Cheers.

Leave a Reply

2 Comments

gravatarTerry

Doesn’t the 1070 have 6+ tflops whereas the 1660 has only 5 tflops? Wouldn’t this make the 1070 better for madvr, especially when the 1070 comes with 8gb of memory?

Reply
gravatarJon Scaife

Possibly so, altho my guess is that the difference isn’t big enough to enable any additional settings in madvr so may be irrelevant. The tflops score doesn’t appear to perfectly predict madvr performance. The only way to know for sure would be to test both but I haven’t found anywhere doing direct comparisons of cards in madvr.

Reply