Leaked slide shows AMD Radeon R9 Nano's great performance per watt

[Scorpus]

It shouldn't be too long before AMD announces the Radeon R9 Nano, but before we see any official details, a leaked slide has appeared that gives us a good indication of what the graphics card will bring in terms of performance.

The leaked image comes courtesy of German website Golem, who obtained the slide from AMD's Hot Chips GPU Session in California. While the slide does contain an obvious spelling error in the title, the performance results shown in the slide look pretty legitimate.

For starters, the R9 Nano's raw performance appears to be slightly ahead of AMD's Hawaii-based Radeon R9 290X in what's apparently a Far Cry 4 benchmark at 4K. More importantly, the R9 Nano smashes both the R9 290X and the R9 Fury X in performance per watt testing, with the Nano recording almost 1.9x the performance per watt of the R9 290X.

Considering the performance figures shown above and the known TDP of the R9 290X (290W), the R9 Nano and its Fiji GPU will likely have a TDP somewhere in the 150 to 175 watt range. This would be a great result for AMD, who have previously struggled to produce high-performance cards with low power consumption.

The R9 Nano is set to be an awesome card for compact system builders, as it features a board just 15cm in length that will fit perfectly into mini-ITX gaming rigs. Considering performance is now suggested to be as good as (or better than) the R9 290X, the R9 Nano could be used for some seriously powerful living room PCs, for example.

AMD is expected to launch the R9 Nano in the coming weeks, so stay tuned for more information.

Permalink to story.

https://www.techspot.com/news/61873-leaked-amd-radeon-r9-nano-performance.html

 

[dividebyzero]

Leaked slide shows AMD Radeon R9 Nano's great performance per watt

Oddly enough, the performance per watt metric is based on a metric never used before to my knowledge.

Power efficiency calculated as GB/s of bandwidth delivered per watt of power consumed

If bandwidth had a direct correlation with performance, I could understand the hype train stoking the boilers, but any cursory check of a comparative graphic performance review will show this not to be the case.

Also pretty odd that AMD chose to use the 290X as a base line considering has been EoL'ed since the 300 series arrived.

 

[amstech]

Never trust data provided by the parent company.

 

[Cryio]

I'm just excited for a review of the Nano.

 

[madboyv1]

I'm just excited for a review of the Nano.

Me too. New small form factor cases and parts (notwithstanding intel's 5x5, I don't quite understand it... yet) get my jimmies all rustled.

 

[Skidmarksdeluxe]

I can't say I get terribly excited about anything AMD releases anymore and it's because of their non existing performance in the performance CPU arena that Intel are able to stroll around with their hands in their pockets and let us squabble over the few crumbs they throw us on occasion.
As for their GPU's, they're not bad performers but they are just too hot and hungry for this day and age but I'll keep an eye on this nano thing of theirs and wouldn't be at all surprised if it was somewhat underwhelming.

 

[Puiu]

They did say before that the Nano will have performance similar to the 290x. This is nothing new.

 

[mrjgriffin]

Amd.....just give me a card that gives me epic fps and doesn't hit 70 degrees celcius unless under VERY intense situations. a card that requires a 5-700 or even greater watt psu is fine with me as long as the performance is there. concentrate on performance man.

 

[wastedkill]

Yay another "performance review" from AMD that looks exactly like the one where the Fury x beat every Nvidia card by a nice margin and that was shown to be a complete lie so this is another lie. This shouldn't be on here same goes for the "1 DX12 game shows Nvidia is now dead" article.

 

[Puiu]

Yay another "performance review" from AMD that looks exactly like the one where the Fury x beat every Nvidia card by a nice margin and that was shown to be a complete lie so this is another lie. This shouldn't be on here same goes for the "1 DX12 game shows Nvidia is now dead" article.

If you set the game settings as the ones used by AMD the Fury X did indeed deliver better performance in many of those games. But without more details it's hard to tell what they benchmark and where.

 

[dividebyzero]

If you set the game settings as the ones used by AMD the Fury X did indeed deliver better performance in many of those games. But without more details it's hard to tell what they benchmark and where.

AMD's "benchmark" settings involved running games without anisotrophic filtering to achieve the overall "result".
In what real scenario does a gamer enable SSAA or 4xMSAA yet disable AF ?

In some other results, carefully chosen settings are enabled at the expense of playabililty - although that particular gambit is certainly a staple of many a PR slide regardless of vendor. It just give the the overall impression of SMH v1.0 when allied with the dodgy 0xAF.

 

[GeforcerFX]

This is the level I think AMD should be focusing more on, they need to take back the Mid level or at least compete there better. I could care less if AMD has the "Fastest card" since a few thousand people buy them, but millions look to buy cards in $150-$300 price point, the re-badge wasn't enough.