580's crossfired or Vega 56

[coaltown992]

So,i'm shopping for a GPU for my first custom build and i just realized crossfire is a thing. I'm wondering if instead of getting a Vega 56, my latest plan, should i just get 2 580's? Also i have a monitor with freesync so i'm pretty set on AMD.

[Lurick]

A single stronger card is, in about 95%+ of cases, the better choice.

I would go with the single Vega 56 if you're set on AMD

[coaltown992]

yeah, that's what i'm leaning towards, this is basically just my last attempt to see if there are any other benefits to cross fire that I haven't heard yet. 

[Streetguru]

15 minutes ago, coaltown992 said:

So,i'm shopping for a GPU for my first custom build and i just realized crossfire is a thing. I'm wondering if instead of getting a Vega 56, my latest plan, should i just get 2 580's? Also i have a monitor with freesync so i'm pretty set on AMD.

Crossfire/SLI is a waste of money, what's your monitor resolution/refresh rate?

[coaltown992]

Just now, Streetguru said:

Crossfire/SLI is a waste of money, what's your monitor resolution/refresh rate?

2560x1080 and 144Hz

[Shimejii]

20 minutes ago, coaltown992 said:

So,i'm shopping for a GPU for my first custom build and i just realized crossfire is a thing. I'm wondering if instead of getting a Vega 56, my latest plan, should i just get 2 580's? Also i have a monitor with freesync so i'm pretty set on AMD.

Crossfire/sli Is HORRID for what you want it to do.

 

I would say get a 1070/1070ti or 1080 for the same price as your going to pay for a Vega 56/64 but if you are so against Nvidia then just be careful with what you find, Quite a few were mining cards.

[Streetguru]

4 minutes ago, coaltown992 said:

2560x1080 and 144Hz

You can probably get away with just a single RX 580 8GB, I assume it's free-sync?

Just drop your shadow quality and AA settings in AAA games.

Save the rest for next gen GPUs in the near future, Nvidia announces some on the 20th most likely.

[Jaggsta]

about $350 with ebay 15% promo for used Vega 56 or $400 for 64

[coaltown992]

47 minutes ago, Jaggsta said:

about $350 with ebay 15% promo for used Vega 56 or $400 for 64

what promo is that? could i get a link?

 

[-rascal-]

11 hours ago, Shimejii said:

Crossfire/sli Is HORRID for what you want it to do.

 

I would say get a 1070/1070ti or 1080 for the same price as your going to pay for a Vega 56/64 but if you are so against Nvidia then just be careful with what you find, Quite a few were mining cards.

 

11 hours ago, coaltown992 said:

So,i'm shopping for a GPU for my first custom build and i just realized crossfire is a thing. I'm wondering if instead of getting a Vega 56, my latest plan, should i just get 2 580's? Also i have a monitor with freesync so i'm pretty set on AMD.

 

If games support multi-GPU configurations, it works pretty good.

The last time I checked profession reviews, two-way GPU scaling was in the 90%+ range.

 

On the other hand, it games DO NOT support mult-GPU configurations, or is poorly optimized, then you get either of the following:

• Second GPU gets unused (sits idle)
• Poor multi-GPU scaling (e.g. +50% with second card)
• Rare cases, NEGATIVE scaling (e.g. drop a few FPS)

 

 

SLi/Crossfire has come a long way.

 

When I did two-way and three-way Crossfire with Radeon HD 5850's, it was like +50% ~ 75% for two cards on average.  On VERY well optimized games, you can see ~90%.

The third card would be +15% ~ 25%. Basically just for bragging rights.

 

By the time I went with two-way Crossfire with Radeon HD 7970's, we were beginning to see 90%+ scaling, up to  near 98~99%, on DX11 titles.

Notable games were BF3, Alien vs. Predator, Crysis 2, etc.

Of course, you still had games that only gained about +50% or +60% with the second GPU.

 

Now I am sitting with two R9-Fury's in Crossfire.

By this time, reviews were showing 75%+ scaling for 1080p and 1440p resolutions.

4K resolutions were seeing 98% to basically 100% scaling.

 

This wasn't just true for AMD cards, nVidia scaled quite well too, like with the GTX 980 and 980 Ti.

 

 

 

All said and done, it is not wroth running SLI or Crossfire on low-end or mid-tier cards.

It would only really be worth it if running on mid-high to high-end GPUs.

In your case, if you decide to run two Vega 56's in Crossfire -- sure.

GTX 1080's or GTX 1080 Ti's? -- Yeah.

GTX 1070's...ehh debatable.

 

[Peskanova]

6 hours ago, Jaggsta said:

about $350 with ebay 15% promo for used Vega 56 or $400 for 64

well I will go for a 56 or a 64 , for second hand ,rx 580 is with 480 the flagship of mining

And despite low tier card 480 (580 believe also) was meant to be in xfire to compete with a 1080 on vr games so its not crazy, you can buy 1 as people said and try 1 a few days (first hand)  and if you are not ok return...  

[Shimejii]

1 minute ago, -rascal- said:

 

 

If games support multi-GPU configurations, it works pretty good.

The last time I checked profession reviews, two-way GPU scaling was in the 90%+ range.

 

On the other hand, it games DO NOT support mult-GPU configurations, or is poorly optimized, then you get either of the following:

• Second GPU gets unused (sits idle)
• Poor multi-GPU scaling (e.g. +50% with second card)
• Rare cases, NEGATIVE scaling (e.g. drop a few FPS)

 

 

SLi/Crossfire has come a long way.

 

When I did two-way and three-way Crossfire with Radeon HD 5850's, it was like +50% ~ 75% for two cards on average.  On VERY well optimized games, you can see ~90%.

The third card would be +15% ~ 25%. Basically just for bragging rights.

 

By the time I went with two-way Crossfire with Radeon HD 7970's, we were beginning to see 90%+ scaling, up to  near 98~99%, on DX11 titles.

Notable games were BF3, Alien vs. Predator, Crysis 2, etc.

Of course, you still had games that only gained about +50% or +60% with the second GPU.

 

Now I am sitting with two R9-Fury's in Crossfire.

By this time, reviews were showing 75%+ scaling, most in the 80% ~ 90% range, for 1080p and 1440p resolutions.

4K resolutions were seeing 98% to basically 100% scaling.

 

This doesn't wasn't just true for AMD cards, nVidia scaled quite well as well, like the GTX 980 and 980 Ti.

 

 

 

All said and done, it is not wroth running SLI or Crossfire on low-end or mid-tier cards.

It would only really be worth it if running on mid-high to high-end GPUs.

In your case, if you decide to run two Vega 56's in Crossfire -- sure.

GTX 1080's or GTX 1080 Ti's? -- Yeah.

GTX 1070's...ehh debatable.

 

Ive personally used 980TI in SLI for a short period (I had one already, built my brother his computer and was simply testing a few things) and sure the numbers LOOK great, but the Input issues, Stuttering, frame drop issues, among the MANY more optimizations, game devs being shit and lazy, its simply NEVER has been worth it nor will it be worth it for gaming. Workstation use and other uses are different, i do not count them since this user stated gaming was the intended use.

 

Running 1080ti in SLI is just dumb, wasteful, and not wo rth the added hassle. Even with 4k 144 hz monitors, the stuttering and other SLI issues would make it not worth it from what ive seen. I would love to see where your Nvidia Numbers are coming from with the Scaling as most ive seen for 90% of the games was UP 70-75% with realistic resolutions and settings in VERY limited games, and in general it was 30-45%

 

SLI Will Never be worth the added cost IMO for any resolution no matter the card with what ive seen from it.

[-rascal-]

3 minutes ago, Shimejii said:

Ive personally used 980TI in SLI for a short period (I had one already, built my brother his computer and was simply testing a few things) and sure the numbers LOOK great, but the Input issues, Stuttering, frame drop issues, among the MANY more optimizations, game devs being shit and lazy, its simply NEVER has been worth it nor will it be worth it for gaming. Workstation use and other uses are different, i do not count them since this user stated gaming was the intended use.

 

Running 1080ti in SLI is just dumb, wasteful, and not wo rth the added hassle. Even with 4k 144 hz monitors, the stuttering and other SLI issues would make it not worth it from what ive seen. I would love to see where your Nvidia Numbers are coming from with the Scaling as most ive seen for 90% of the games was UP 70-75% with realistic resolutions and settings in VERY limited games, and in general it was 30-45%

 

SLI Will Never be worth the added cost IMO for any resolution no matter the card with what ive seen from it.

 

True, there are games that scale like garbage.

That said, there are games out there that scales pretty good as well.

 

If you are pushing for higher resolutions, meaning beyond 1080p and 1440p, one card would not be enough, considering you see people in the LTT community who demands for as high graphical settings as possible. Multi-GPU configurations are a mixed bag, and I get it.

 

https://www.guru3d.com/articles-pages/geforce-gtx-980-sli-review,1.html

Just a few from their review article..note they test with 2-way and 3-way SLI

DX11 Hitman Absolution:

• 2560x1440 -- 65 FPS (1 GPU), 101 FPS (2 GPUs) +50%,

DX11 Tomb Raider:

• 1920x1200 -- 146 FPS (1 GPU), 262 FPS (2 GPUS) +80%
• 2560x1440 -- 100 FPS (1 GPU), 187 FPS (2 GPUs) +87%

DX11 Medal of Honor Warfighter

• 1920x1200 -- 96 FPS (1 GPU), 176 FPS (2 GPUs) +84%
• 2560x1440 -- 65 FPS (1 GPU), 122 FPS (2 GPUs) +88%

DX11 Crysis 3

• 1920x1200 -- 67 FPS (1 GPU), 109 FPS (2 GPUs) +65%
• 2560x1440 -- 48 FPS (1 GPU), 86 FPS (2 GPUs) +80%

 

 

https://www.overclock3d.net/reviews/gpu_displays/asus_gtx980_ti_matrix_sli_review/1

Battlefield Hardline

• 1920x1080 -- 124.9 FPS (1 GPU), 199.9 FPS (2 GPUs) +60%
• 2560x1440 -- 78.5 FPS (1 GPU), 157 FPS (2 GPUs) +100%
• 3840x2160 -- 44.6 FPS (1 GPU), 80.7 FPS (2 GPUs) +80%

Metro 2033

• 1920x1080 -- 94.8 FPS (1 GPU), 173.2 FPS (2 GPUs) +83%
• 2560x1440 -- 60.4 FPS (1 GPU), 107.1 FPS (2 GPUs) +77%
• 3840x2160 -- 25.3 FPS (1 GPU), 49.4 FPS (2 GPUs) +95%

Sleeping Dogs

• 1920x1080 -- 122.4 FPS (1 GPU), 204.3 FPS (2 GPUs) +67%
• 2560x1440 -- 74.1 FPS (1 GPU), 139.8 FPS (2 GPUs) +87%
• 3840x2160 -- 32.7 FPS (1 GPU), 63.9 FPS (2 GPUs) +96%

 

 https://www.gamersnexus.net/guides/1970-gtx-980-ti-sli-benchmark-vs-sli-980-titan-x/Page-2

Witcher 3 @ 1440p

GTX 980 -- 33 FPS 0.1%, 37 FPS 1%, 45 FPS Average

GTX 980 SLI -- 42 FPS 0.1%, 51 FPS 1%, 77 FPS Average

GTX 980 Ti -- 41 FPS 0.1%, 45 FPS 1%, 57 FPS Average

GTX 980 TI SLI -- 53 FPS 0.1%, 60 FPS 1%,  86 FPS Average

 

Shadow of Mordor @ 1440p

GTX 980 -- 42 FPS 0.1%, 48 FPS 1%, 62 FPS Average

GTX 980 SLI -- 71 FPS 0.1%, 82 FPS 1%, 109 FPS Average

GTX 980 Ti -- 60 FPS 0.1%, 64 FPS 1%, 79 FPS Average

GTX 980 TI SLI -- 84 FPS 0.1%, 92 FPS 1%,  130 FPS Average

 

Far Cry 4 @ 4K

GTX 980 -- 22 FPS 0.1%, 29 FPS 1%, 35 FPS Average

GTX 980 SLI -- 39 FPS 0.1%, 58 FPS 1%, 84 FPS Average

GTX 980 Ti -- 27 FPS 0.1%, 36 FPS 1%, 43 FPS Average

GTX 980 TI SLI -- 39 FPS 0.1%, 58 FPS 1%,  84 FPS Average

 

Far Cry 4 @ 1440p

GTX 980 -- 35 FPS 0.1%, 47 FPS 1%, 63 FPS Average

GTX 980 SLI -- 38 FPS 0.1%, 63 FPS 1%, 103 FPS Average

GTX 980 Ti -- 43 FPS 0.1%, 59 FPS 1%, 78 FPS Average

GTX 980 TI SLI -- 39 FPS 0.1%, 71 FPS 1%,  109 FPS Average

[Shimejii]

19 minutes ago, -rascal- said:

 

True, there are games that scale like garbage.

That said, there are games out there that scales pretty good as well.

 

If you are pushing for higher resolutions, meaning beyond 1080p and 1440p, one card would not be enough, considering you see people in the LTT community who demands for as high graphical settings as possible. Multi-GPU configurations are a mixed bag, and I get it.

 

https://www.guru3d.com/articles-pages/geforce-gtx-980-sli-review,1.html

Just a few from their review article..note they test with 2-way and 3-way SLI

DX11 Hitman Absolution:

• 2560x1440 -- 65 FPS (1 GPU), 101 FPS (2 GPUs) +50%,

DX11 Tomb Raider:

• 1920x1200 -- 146 FPS (1 GPU), 262 FPS (2 GPUS) +80%
• 2560x1440 -- 100 FPS (1 GPU), 187 FPS (2 GPUs) +87%

DX11 Medal of Honor Warfighter

• 1920x1200 -- 96 FPS (1 GPU), 176 FPS (2 GPUs) +84%
• 2560x1440 -- 65 FPS (1 GPU), 122 FPS (2 GPUs) +88%

DX11 Crysis 3

• 1920x1200 -- 67 FPS (1 GPU), 109 FPS (2 GPUs) +65%
• 2560x1440 -- 48 FPS (1 GPU), 86 FPS (2 GPUs) +80%

 

 

https://www.overclock3d.net/reviews/gpu_displays/asus_gtx980_ti_matrix_sli_review/1

Battlefield Hardline

• 1920x1080 -- 124.9 FPS (1 GPU), 199.9 FPS (2 GPUs) +60%
• 2560x1440 -- 78.5 FPS (1 GPU), 157 FPS (2 GPUs) +100%
• 3840x2160 -- 44.6 FPS (1 GPU), 80.7 FPS (2 GPUs) +80%

Metro 2033

• 1920x1080 -- 94.8 FPS (1 GPU), 173.2 FPS (2 GPUs) +83%
• 2560x1440 -- 60.4 FPS (1 GPU), 107.1 FPS (2 GPUs) +77%
• 3840x2160 -- 25.3 FPS (1 GPU), 49.4 FPS (2 GPUs) +95%

Sleeping Dogs

• 1920x1080 -- 122.4 FPS (1 GPU), 204.3 FPS (2 GPUs) +67%
• 2560x1440 -- 74.1 FPS (1 GPU), 139.8 FPS (2 GPUs) +87%
• 3840x2160 -- 32.7 FPS (1 GPU), 63.9 FPS (2 GPUs) +96%

 

 https://www.gamersnexus.net/guides/1970-gtx-980-ti-sli-benchmark-vs-sli-980-titan-x/Page-2

Witcher 3 @ 1440p

GTX 980 -- 33 FPS 0.1%, 37 FPS 1%, 45 FPS Average

GTX 980 SLI -- 42 FPS 0.1%, 51 FPS 1%, 77 FPS Average

GTX 980 Ti -- 41 FPS 0.1%, 45 FPS 1%, 57 FPS Average

GTX 980 TI SLI -- 53 FPS 0.1%, 60 FPS 1%,  86 FPS Average

 

Shadow of Mordor @ 1440p

GTX 980 -- 42 FPS 0.1%, 48 FPS 1%, 62 FPS Average

GTX 980 SLI -- 71 FPS 0.1%, 82 FPS 1%, 109 FPS Average

GTX 980 Ti -- 60 FPS 0.1%, 64 FPS 1%, 79 FPS Average

GTX 980 TI SLI -- 84 FPS 0.1%, 92 FPS 1%,  130 FPS Average

 

Far Cry 4 @ 4K

GTX 980 -- 22 FPS 0.1%, 29 FPS 1%, 35 FPS Average

GTX 980 SLI -- 39 FPS 0.1%, 58 FPS 1%, 84 FPS Average

GTX 980 Ti -- 27 FPS 0.1%, 36 FPS 1%, 43 FPS Average

GTX 980 TI SLI -- 39 FPS 0.1%, 58 FPS 1%,  84 FPS Average

 

Far Cry 4 @ 1440p

GTX 980 -- 35 FPS 0.1%, 47 FPS 1%, 63 FPS Average

GTX 980 SLI -- 38 FPS 0.1%, 63 FPS 1%, 103 FPS Average

GTX 980 Ti -- 43 FPS 0.1%, 59 FPS 1%, 78 FPS Average

GTX 980 TI SLI -- 39 FPS 0.1%, 71 FPS 1%,  109 FPS Average

I know Steve has mentioned the Stuttering and other issues with SLI, although im surprised that they never mentioned it in either article. Unless they purely didnt show in benchmarking. I use the X99 5930k @4.4 ghz when i did my testing with it, so would be similar enough. Although who the fuck plays at 1920x1200? 

[Leanora]

14 hours ago, coaltown992 said:

So,i'm shopping for a GPU for my first custom build and i just realized crossfire is a thing. I'm wondering if instead of getting a Vega 56, my latest plan, should i just get 2 580's? Also i have a monitor with freesync so i'm pretty set on AMD.

Get a single card, crossfire isn't worth the headache. 

As for vega 56, depending on local prices a 64 might even be cheaper to get.