CF RX 480

[Leyawiin]

links ??



I was comparing a single GTX 1070 to CF RX 470, the Single GTX 1070 has no need for CF/SLI profiles and Scaling. Its one of the pros for the Single GTX 1070 vs CF RX 470.

Still, whether the CF RX 470 is faster is debatable and and at best situational. As usual the recommendation of a single, stronger card holds. I would personally sell the single RX 470 if I had one and apply it to the purchase of a GTX 1070 vs. buying another to CF. 4GB more VRAM, no multi-GPU issues/more consistent performance, 150W less power usage - seems worth the extra cost to me.

http://www.gamersnexus.net/guides/2547-amd-rx-470-crossfire-benchmark-vs-rx-480-and-1070
http://www.eteknix.com/amd-rx-470-crossfire-graphics-card-review/13/
http://lanoc.org/review/video-cards/7309-rx-470-crossfire-testing?showall=&start=2

 

[bystander36]

Today's prices from newegg,

RX 470 4GB = $150 AR

2x RX 470 4GB = $300 AR

Cheapest GTX1070 = $385 AR

CF RX 470 vs GTX 1070

CF = cheaper
CF = Faster

GTX 1070 lower consumption
GTX 1070 no CF profile/scaling

Its a nice alternative, better to have it than not. Its not for everybody but some may find it worth it to use it.

@guskline

Could you please report after the new driver is officially launched if Chill can be used in CF mode ??

I think one of the issues you have here, is the rx 480 CF is barely faster than a GTX 1070. A slower card in CF is not.

And no one ever seems to note that in CF/SLI, the same FPS does not feel as smooth as a single card. At least for me, who gets nauseated easily by latency and lack of smoothness.

 

[tential]

Still, whether the CF RX 470 is faster is debatable and and at best situational. As usual the recommendation of a single, stronger card holds. I would personally sell the single RX 470 if I had one and apply it to the purchase of a GTX 1070 vs. buying another to CF. 4GB more VRAM, no multi-GPU issues/more consistent performance, 150W less power usage - seems worth the extra cost to me.

http://www.gamersnexus.net/guides/2547-amd-rx-470-crossfire-benchmark-vs-rx-480-and-1070
http://www.eteknix.com/amd-rx-470-crossfire-graphics-card-review/13/
http://lanoc.org/review/video-cards/7309-rx-470-crossfire-testing?showall=&start=2

This is the review I've been referencing. Mind linking some of the charts to the thread since I'm on mobile and I doubt people will click into them.

I only brought cf 470 by the way since I think the cf 480 price to performance is way too poor but the 470s isn't bad.

This goes straight to the post a couple posts up stating people are too black and white. There are situations where you'd want cf over a 1070 and situations where you'd want a titan xp. It's so dependent on the person and their games played and goals.

 

[happy medium]

This is the review I've been referencing

I don't see the link??

edit:

this one?

 

[AtenRa]

A single 1070 is 90% faster than a 4GB 470 so there is really no need for a link here...

Almost double the performance ??? links ???

 

[Thinker_145]

Also keep in mind that the difference will actually get bigger in future due to VRAM bottleneck.

 

[happy medium]

Almost double the performance ??? links ???

"than a" 470. I think you missed that.

 

[happy medium]

Also keep in mind that the difference will actually get bigger in future due to VRAM bottleneck.

85% is close enough.

 

[happy medium]

Almost double the performance ??? links ???

 

[AtenRa]

any other links except TPU ???

Have you seen this one ???

http://lanoc.org/review/video-cards/7309-rx-470-crossfire-testing?showall=&start=2

They used 16.7.3 since this review is from August. With latest drivers and Game patches the performance of single RX 470 and CF is higher today than in August.

 

[Thinker_145]

85% is close enough.

No that's actually 90% a lot of people make that mathematical error.

179-94 = 85

But that's not the actual number. 85/94 x 100 would give you the percentage which comes to 90% in this case.

 

[Thinker_145]

any other links except TPU ???

Have you seen this one ???

http://lanoc.org/review/video-cards/7309-rx-470-crossfire-testing?showall=&start=2

They used 16.7.3 since this review is from August. With latest drivers and Game patches the performance of single RX 470 and CF is higher today than in August.

Sorry for using quite an old review. Happy medium posted a recent review from TPU and the difference we see is 83% which still leaves no room for debate on whether 2 such cards can compete with 1.

 

[AtenRa]

Latest 2016 games

RX 470 4GB = 59fps
GTX 1070 = 83fps

GTX 1070 = ~40% faster in BF1

RX 470 4GB = 49fps
GTX 1070 = 78fps

GTX 1070 = ~59% faster in Gears of War 4

RX 470 4GB = 46fps
GTX 1070 = 66fps

GTX 1070 = ~43% faster in Titanfall 2

RX 470 4GB = 34fps
GTX 1070 = 58fps

GTX 1070 = ~70% faster in Dishonored 2

RX 470 4GB = 66ps
GTX 1070 = 97fps

GTX 1070 = ~47% faster in Call of Duty

RX 470 4GB = 28ps
GTX 1070 = 47fps

GTX 1070 = ~68% faster in Watch Dogs 2

RX 470 4GB = 49ps
GTX 1070 = 80fps

GTX 1070 = ~63% faster in Civ VI

RX 470 4GB = 28ps
GTX 1070 = 37fps

GTX 1070 = ~32% faster in Dues Ex MD

The average of those 8 Games is close to 53%

 

[happy medium]

any other links except TPU ???

Have you seen this one ???

http://lanoc.org/review/video-cards/7309-rx-470-crossfire-testing?showall=&start=2

They used 16.7.3 since this review is from August. With latest drivers and Game patches the performance of single RX 470 and CF is higher today than in August.

http://www.gamersnexus.net/guides/2547-amd-rx-470-crossfire-benchmark-vs-rx-480-and-1070

http://www.eteknix.com/amd-rx-470-crossfire-graphics-card-review/13/

 

[happy medium]

Latest 2016 games

RX 470 4GB = 59fps
GTX 1070 = 83fps

GTX 1070 = ~40% faster in BF1

RX 470 4GB = 49fps
GTX 1070 = 78fps

GTX 1070 = ~59% faster in Gears of War 4

RX 470 4GB = 46fps
GTX 1070 = 66fps

GTX 1070 = ~43% faster in Titanfall 2

RX 470 4GB = 34fps
GTX 1070 = 58fps

GTX 1070 = ~70% faster in Dishonored 2

RX 470 4GB = 66ps
GTX 1070 = 97fps

GTX 1070 = ~47% faster in Call of Duty

RX 470 4GB = 28ps
GTX 1070 = 47fps

GTX 1070 = ~68% faster in Watch Dogs 2

RX 470 4GB = 49ps
GTX 1070 = 80fps

GTX 1070 = ~63% faster in Civ VI

RX 470 4GB = 28ps
GTX 1070 = 37fps

GTX 1070 = ~32% faster in Dues Ex MD

The average of those 8 Games is close to 53%

God man do you really have to cherry pick benchmarks all the time?
This is not my job, I don't have time to pick the other 30 benchmarks that show your wrong.
When your wrong, your wrong, give it up.

 

[happy medium]

anyway thanks for the 480 x fire benchmarks Guskline. I'm out.

 

[AtenRa]

God man do you really have to cherry pick benchmarks all the time?

I have just posted 8 of the latest 2016 games, if you believe those are cherry picking thats fine but i dont. And you also only provided a single source, some people could also say you are cherry picking with posting a single TPU graph.

Latest TPU reviews doesnt include all of the cards in the separated Game graphs. just to point out that we have no numbers how RX 470 performs in Civilization VI for example.

 

[Thinker_145]

lol nice cherry picking there. The TPU review includes 11 games from 2016 and 10 from 2015. That's the way it works. All these best case scenarios that you have posted are taken into account but the average still comes to 83% which says a thing or 2 about the worst case scenarios.

 

[Rifter]

God man do you really have to cherry pick benchmarks all the time?
This is not my job, I don't have time to pick the other 30 benchmarks that show your wrong.
When your wrong, your wrong, give it up.

How is cherry picking benchmarks any different than posting almost 6 months old TWP charts that have no bearing on current performance with current drivers.

I agree that the stronger single card is better if you need the performance, i just wish people would at least use current benchmarks when trying to make a case.

 

[AtenRa]

lol nice cherry picking there. The TPU review includes 11 games from 2016 and 10 from 2015. That's the way it works. All these best case scenarios that you have posted are taken into account but the average still comes to 83% which says a thing or 2 about the worst case scenarios.

could you post the graph from the TPU review of individual games ??

Example, Civilization VI that has both the RX 470 and GTX 1070, or any other game

 

[Thinker_145]

The RX 470 and 1070 are too far apart to feature in individual games in a TPU review as they only show relevant cards to the one being reviewed. But at the end of the review when you see the overall numbers all cards were run for all the games listed in the review with the latest drivers.

 

[tential]

The RX 470 and 1070 are too far apart to feature in individual games in a TPU review as they only show relevant cards to the one being reviewed. But at the end of the review when you see the overall numbers all cards were run for all the games listed in the review with the latest drivers.

Do we have a source for this? I had been told that tpu reuses numbers and doesn't retest all gpus for each review. I mean, if you do a 1050ti review, and don't include the fury x results or whatever high end gpu in the individual charts... Why not? I mean if you tested every for why not include all data?


We should be careful though and make sure this is the case.

 

[Thinker_145]

Do we have a source for this? I had been told that tpu reuses numbers and doesn't retest all gpus for each review. I mean, if you do a 1050ti review, and don't include the fury x results or whatever high end gpu in the individual charts... Why not? I mean if you tested every for why not include all data?


We should be careful though and make sure this is the case.

Because it's not relevant to the review in question. A smaller chart keeps the information easier to use.

Yes it has been stated multiple times by them in the past that they retest with every card whenever a new driver is released. You think they would have a shred of credibility if they lumped overall performance charts using different drivers?

 

[tential]

Because it's not relevant to the review in question. A smaller chart keeps the information easier to use.

Yes it has been stated multiple times by them in the past that they retest with every card whenever a new driver is released. You think they would have a shred of credibility if they lumped overall performance charts using different drivers?

Ok, so they don't retest every review, just every driver update. Gotcha, that's probably what I saw

 

[Thinker_145]

Ok, so they don't retest every review, just every driver update. Gotcha, that's probably what I saw

They retest whenever a variable changes like a new game version. No point retesting when nothing has changed.