[Rumor, Tweaktown] AMD to launch next-gen Navi graphics cards at E3

[Glo.]

One rumor has it that they will be priced ~RX570-580 MSRP ($169-229), yet another reports $149 for 5500 and $199 for 5500 XT.

If they put out a 5500 at $149 and XT at $199 and they scale performance well from 5700/XT, I'll need a change of underpants.

199$ for XT would only make sense if it was full die. 149$ for RX 5500 4 GB version? I can see this happening.

 

[lifeblood]

One rumor has it that they will be priced ~RX570-580 MSRP ($169-229), yet another reports $149 for 5500 and $199 for 5500 XT.

If they put out a 5500 at $149 and XT at $199 and they scale performance well from 5700/XT, I'll need a change of underpants.

The 5500 (vanilla) performs between the 570 & 580. The 570 sells for around $139 with the 580 around $179. Selling the 5500 for $10-$30 more than the 570 does us no favors. Of course that could be useful for helping to clear out existing stocks of the 570, then they could drop the price a bit later.

Also, no benchmarks. All we got is AMD's marketing talk. I really want to see independent benchmarks of the retail cards before I make any definitive conclusions about price and such.

Either way this does me no good. Looks like I'm going to end up with a 1660 Ti. Not that that's a bad thing, the 1660 Ti is a good card. It just would have been nice to have a choice. Maybe a reference 5700 will be $250 on Black Friday. I'd absolutely jump on that over a 1660 Ti.

*Edited for grammar/missed words

 

[amrnuke]

The 5500 (vanilla) performs between the 570 & 580. The 570 sells for around $139 with the 580 around $179. Selling the 5500 for $10-$30 more than the 570 does us no favors. Of course that could be useful for helping to clear out existing stocks of the 570, then they could drop the price a bit later.

Also, no benchmarks. All we got is AMD's marketing talk. I really want to see independent benchmarks of the retail cards before I make any definitive conclusions about price and such.

Either way this does me no good. Looks like I'm going to end up with a 1660 Ti. Not that that's a bad thing, the 1660 Ti is a good card. It just would have been to have a choice. Maybe a reference 5700 will be $250 on Black Friday. I'd absolutely jump on that over a 1660 Ti.

You said the 5500 performs between the 570 and 580, but then said "all we got is AMD's marketing talk" and "no benchmarks". So which is it? You can't make a proclamation that the 5500 and XT at $149/199 "does us no favors" when we don't know the performance. Or maybe you do, as you stated in your first sentence?

Rather than saying "it just would have been [nice] to have a choice" and dismissing these cards, why not wait for benchmarks and see whether they offer a choice?

All indications are that the 5500/XT will be 1660/1660S/1660Ti level performance, and it appears possibly at $149-$229 price range.

 

[Glo.]

The 5500 (vanilla) performs between the 570 & 580. The 570 sells for around $139 with the 580 around $179. Selling the 5500 for $10-$30 more than the 570 does us no favors. Of course that could be useful for helping to clear out existing stocks of the 570, then they could drop the price a bit later.

Also, no benchmarks. All we got is AMD's marketing talk. I really want to see independent benchmarks of the retail cards before I make any definitive conclusions about price and such.

Either way this does me no good. Looks like I'm going to end up with a 1660 Ti. Not that that's a bad thing, the 1660 Ti is a good card. It just would have been to have a choice. Maybe a reference 5700 will be $250 on Black Friday. I'd absolutely jump on that over a 1660 Ti.

The data points to the fact that vanilla RX 5500 being just slightly behind GTX 1660, not between RX 580 and 570.

I have no idea what you base on your assumptions.

Lets take worst case scenario, for Overwatch.

1080p, Epic settings. GTX 1060 averages 125 FPS, RX 580 125, RX 590 128 FPS in those settings, with Intel CPU. All of them have to be paired with Intel CPU here.

AMD tested their GPUs with Ryzen 3 3600X, and it averaged 135 FPS in those settings.

You know what this means? With Intel CPU AMD GPUs are around 10-15 FPS FASTER than they are with Ryzen CPUs, in this specific game. So EXACTLY in the ballpark of 145-150 FPS.

GTX 1660 averages with Intel CPU in this game 144 FPS, while GTX 1660 Ti averages 155 FPS in this game with Intel CPU.

This GPU is NOT on the same level as RX 570. Put that out of your heads guys, we are talking about GTX 1660-1660 Ti performance for much less money.

 

[lifeblood]

The data points to the fact that vanilla RX 5500 being just slightly behind GTX 1660, not between RX 580 and 570.

I have no idea what you base on your assumptions.

According to the few articles I've read AMD claims it performs around the RX 480. The 480 performs between a 570 and a 580. Again the 5500 XT (assuming their is one) should be around the 590 level of performance. Also, AMD is saying the 5500 is a GTX 1650 competitor rather than 1660. Their is a reason they're saying that. Maybe the 5500 has 1660 performance at 1650 prices, that would be awesome, but something tells me that's not it. Again, benchmarks and official prices will tell the whole story.

Do we have benchmarks prices out yet? I haven't seen any. That's why I said I would like to see them. We could stop guessing based on estimates and argue hard numbers.

 

[Glo.]

According to the few articles I've read AMD claims it performs around the RX 480. The 480 performs between a 570 and a 580. Again the 5500 XT (assuming their is one) should be around the 590 level of performance. Also, AMD is saying the 5500 is a GTX 1650 competitor rather than 1660. Their is a reason they're saying that. Maybe the 5500 has 1660 performance at 1650 prices, that would be awesome, but something tells me that's not it. Again, benchmarks and official prices will tell the whole story.

Do we have benchmarks prices out yet? I haven't seen any. That's why I said I would like to see them. We could stop guessing based on estimates and argue hard numbers.

Those articles said that it OUTPERFORMS RX 480, not performs like RX 480.

And AMD with RX 5700 XT release Undersold the GPU. Why the story has to be different this time?

Benchmarks AMD provided, with specs.

 

[Glo.]

Update, AMD changed the TBP specs for this GPU, and now it states... 150W, instead of 110W.

 

[AtenRa]

According to videocardz, NAVI 14 is 6.4B transistors and that is extremely close to 6.6B of the TU116 (GTX1660/Ti). For reference TU117 (GTX1650) is only 4.7B transistors.

Since NAVI 10 (RX5700/XT) is also extremely close to TU106 (RTX2060/Super) , I would say that RX5500 could be very close to GTX1660 and RX5500XT very close to GTX1660Ti.

Now about the price, since we now that NVIDIA preparing a GTX1650Ti , I believe that RX5500 4GB MSRP could fall between GTX1650 and GTX1660.

 

[Veradun]

According to the few articles I've read AMD claims it performs around the RX 480.

This is 5500 4G

 

[amrnuke]

This is 5500 4G

But if you look at the system setups they used... it is very deceiving.

 

[amrnuke]

Update, AMD changed the TBP specs for this GPU, and now it states... 150W, instead of 110W.

Is this for the XT? The 5700 only pulls 180W under peak gaming load.

If this is for the 5500XT, then this is good news (performance actually could be on-par with 1660Ti though may be less efficient) but also weird news (why isn't efficiency scaling?). The price will be the most important part.

 

[Glo.]

According to videocardz, NAVI 14 is 6.4B transistors and that is extremely close to 6.6B of the TU116 (GTX1660/Ti). For reference TU117 (GTX1650) is only 4.7B transistors.

Since NAVI 10 (RX5700/XT) is also extremely close to TU106 (RTX2060/Super) , I would say that RX5500 could be very close to GTX1660 and RX5500XT very close to GTX1660Ti.

Now about the price, since we now that NVIDIA preparing a GTX1650Ti , I believe that RX5500 4GB MSRP could fall between GTX1650 and GTX1660.

It is close to TU116 .

Lets take Overwatch as an example.

AMD tested RX 5500 with RX 3600X CPU, and 3200 MHz RAM. What this means?

With the same config, GTX 1660 Ti averages 140 FPS in Overwatch 1080p, Epic presets, On some maps its higher, on some maps its lower, on some maps it nails this point. Overall - average it is 140 FPS.

RX 5500 in the same settings, same preset, same CPU, averaged, according to AMD 135 FPS.

You canexpect that with Intel CPU, both of those GPU will get 150 FPS, or higher(GTX 1660 Ti averages actually 155 FPS in Overwatch, 1080p, Epic preset with 9400F CPU).

AMD undersold their product, once again .

 

[amrnuke]

It is close to TU116 .

Lets take Overwatch as an example.

AMD tested RX 5500 with RX 3600X CPU, and 3200 MHz RAM. What this means?

With the same config, GTX 1660 Ti averages 140 FPS in Overwatch 1080p, Epic presets, On some maps its higher, on some maps its lower, on some maps it nails this point. Overall - average it is 140 FPS.

RX 5500 in the same settings, same preset, same CPU, averaged, according to AMD 135 FPS.

You canexpect that with Intel CPU, both of those GPU will get 150 FPS, or higher(GTX 1660 Ti averages actually 155 FPS in Overwatch, 1080p, Epic preset with 9400F CPU).

AMD undersold their product, once again .

If the RX 5500 comes within 5% of the 1660Ti, and there's still an XT version out there... and if the pricing is as rumored... wow.

 

[Glo.]

Interesting. Slide deck says that RX 5500 DT series has 1717 MHz game clock, with boost clock at 1845 MHz.

While 4 GB RX 5500 has 1670 MHz average game clock...

So the higher game clock reserved for the RX 5500 XT?

AMD has done typical AMD release. Absolutely terrible communication. But I think it will be relatively soon be known why they have made it this badly. Nvidia soon launches new GPUs.

 

[AtenRa]

But if you look at the system setups they used... it is very deceiving.

Any links for the entire presentation deck ??

 

[AtenRa]

If the RX 5500 comes within 5% of the 1660Ti, and there's still an XT version out there... and if the pricing is as rumored... wow.

Well, according to this slide (techpowerup) the RX5500 is 12% faster than RX480, witch means that it could be between RX580 and RX590 performance. That will put it approximately within 10% of GTX1660 and not GTX1660Ti.

 

[Glo.]

AtenRa - AMD being AMD - underselling their product .

 

[AtenRa]

WTF ????

The slide says up to 37% faster but if you look at the Overwatch numbers the RX5500 is ~52% faster than the GTX1650.

 

[amrnuke]

Any links for the entire presentation deck ??

It's just straight up on AMD's RX 5500 page.

Footnote 6:
"Testing done by AMD performance labs on August 29, 2019. Systems tested were: Radeon RX 5500 4GB with Ryzen 7 3800X. 16GB DDR4-3200MHz Win10 Pro x64 18362.175. AMD Driver Version 19.30-190812n Vs Radeon RX 480 8GB with Core i7-5960X (3.0GHz) 16GB DDR4-2666 MHz Win10 14393 AMD Driver version 16.10.1 The RX 5500 graphics card provides 1.6x performance per watt, and up to 1.7X performance per area compared to Radeon™ RX 480 graphics. PC manufacturers may vary configurations yielding different results. Actual performance may vary. RX-382"

 

[Elfear]

WTF ????

The slide says up to 37% faster but if you look at the Overwatch numbers the RX5500 is ~52% faster than the GTX1650.

"Up to 37% faster expected AVERAGE performance" The slides show some individual games that are better than a 37% increase (e.g. Overwatch and Rainbow Six) and some that are lower than 37% (e.g. Fortnite & PUBG). They're just talking the average expected increase.

 

[AtenRa]

It's just straight up on AMD's RX 5500 page.

Footnote 6:
"Testing done by AMD performance labs on August 29, 2019. Systems tested were: Radeon RX 5500 4GB with Ryzen 7 3800X. 16GB DDR4-3200MHz Win10 Pro x64 18362.175. AMD Driver Version 19.30-190812n Vs Radeon RX 480 8GB with Core i7-5960X (3.0GHz) 16GB DDR4-2666 MHz Win10 14393 AMD Driver version 16.10.1 The RX 5500 graphics card provides 1.6x performance per watt, and up to 1.7X performance per area compared to Radeon™ RX 480 graphics. PC manufacturers may vary configurations yielding different results. Actual performance may vary. RX-382"

Axaxaxaxa those PR guys again messing with numbers

They compared card performance divided by cards TDP

RX480 = 100% performance and 150W TDP
RX5500 = 112% performance and 105W TDP

100 / 150 = 0.66
112 / 105 = 1.06

1.06 / 0.66 = 1.6X better perf/watt

 

[AtenRa]

"Up to 37% faster expected AVERAGE performance" The slides show some individual games that are better than a 37% increase (e.g. Overwatch and Rainbow Six) and some that are lower than 37% (e.g. Fortnite & PUBG). They're just talking the average expected increase.

Ahh my bad, somehow i missed that

Thanks

edit: Ok now im confused, if we use Average performance why we use the Up-To as well, its not needed correct ??

 

[Glo.]

In terms of sending a message, we can give AMD a prize for "The most unfathomable PR releases" of 2019.

 

[amrnuke]

Axaxaxaxa those PR guys again messing with numbers

They compared card performance divided by cards TDP

RX480 = 100% performance and 150W TDP
RX5500 = 112% performance and 105W TDP

100 / 150 = 0.66
112 / 105 = 1.06

1.06 / 0.66 = 1.6X better perf/watt

Seriously... it's like... what are they even doing?

I don't care how a 5500 + 3800X (3.9/4.5GHz) with DDR4-3200 RAM and the latest drivers performs compared to a 480 + i7-5960X (3.0/3.5GHz) with DDR4-2666 RAM and old drivers, regardless of wattage!

 

[lifeblood]

Seriously... it's like... what are they even doing?

I don't care how a 5500 + 3800X (3.9/4.5GHz) with DDR4-3200 RAM and the latest drivers performs compared to a 480 + i7-5960X (3.0/3.5GHz) with DDR4-2666 RAM and old drivers, regardless of wattage!

And this is why it's wise to look at AMD's marketing claims with a bit of cynicism. When comparing two GPU's you put them in identical PC's. What AMD is doing is comparing system to system, not GPU to GPU.

The 5500 (non XT) will perform just below the 1660 (in identically configured systems and I shouldn't have to specify that). Hopefully it will be priced around the 1650. The 5500 XT, if it ever exists, should perform near the 1660 Ti. And if the 5500 (non XT) ends up beating the 1660 while costing less then I will humbly apologize.

 

[Glo.]

And this is why it's wise to look at AMD's marketing claims with a bit of cynicism. When comparing two GPU's you put them in identical PC's. What AMD is doing is comparing system to system, not GPU to GPU.

The 5500 (non XT) will perform just below the 1660 (in identically configured systems and I shouldn't have to specify that). Hopefully it will be priced around the 1650. The 5500 XT, if it ever exists, should perform near the 1660 Ti. And if the 5500 (non XT) ends up beating the 1660 while costing less then I will humbly apologize.

You once again completely ignore benchmarks AMD released, and their configs, that have been tested on.

135 FPS in Overwatch 1080p, Epic settings, with R5 3600X. GTX 1660 Ti in the same settings, and with the same CPU averages 140 FPS.

This GPU will not be below GTX 1660. It will be slightly below GTX 1660 Ti, and this is the reason why Nvidia is readying GTX 1660 Super, because standard GTX 1660 cannot compete with this GPU.

 

[amrnuke]

You once again completely ignore benchmarks AMD released, and their configs, that have been tested on.

135 FPS in Overwatch 1080p, Epic settings, with R5 3600X. GTX 1660 Ti in the same settings, and with the same CPU averages 140 FPS.

This GPU will not be below GTX 1660. It will be slightly below GTX 1660 Ti, and this is the reason why Nvidia is readying GTX 1660 Super, because standard GTX 1660 cannot compete with this GPU.

What were the RAM timings? Was the 3600X in the 1660Ti rig running PBO or stock? Where did you find the data on endnote RX-383? I thought they ran those 5500 numbers on a 3800X.