Review AMD Radeon RX 9070 XT and RX 9070 review: An excellent value, if supply is good

[JarredWaltonGPU]

I may be off base here as my world is mostly virtualized servers and VDI but... Is a WDS/WDT server holding preconfigured images with all the games and benchmarking software already installed and activated not possible? Have sysprepped images been attempted in the past and they just don't work?

It would "work," but imaging still takes time and the images would be quickly outdated. I probably download at least 200GB of updates to games in my test suite every month. Sometimes more. So if I make an image today, I will still have 30 minutes or whatever of downloading updates once I put the image back in place.

And that's not accounting for the time it takes to actually reimage the drive — about 3TB of data, so even at ~1 GB/s that's nearly an hour. (That 1 GB/s figure is just an estimate based on SSD speeds and the imaging tools, as you rarely would get full write speeds of ~3 GB/s.)

Realistically, if I reimage between every GPU swap, I'm adding at least two hours to testing. And people already complain that "you didn't include the RTX 4070 Super because you're trying to hide things!" Yeah, right. It's not that I'm one person with one testbed doing all the work. It's that I'm trying to hide stuff! LOL

 

[DS426]

...

I think that you meant to say that the RX 9070 XT is faster than the RX 7900 XTX in RT, not raster. The RX 7900 XTX is still the king of the Radeon hill when it comes to raster performance, being 10% faster than the RX 9070 XT on average. When it comes to RT, the RX 9070 XT leaves the RX 7900 XTX in its dust. I own an RX 7900 XTX and I'm perfectly ok with its (lack of) RT performance because don't care about RT, like, at all. I'm also perfectly happy with FSR3.x. I haven't had to use it yet but I still tried it out just to see what it's like and it's more than good enough for me if I ever have to use it. I tried it out in Starfield and had actually played it for a week with FSR enabled without realising it. That tells me it's definitely good enough for my purposes.
...

Oops, yes, I did mean to say it that way, lol. I'll go back and edit that post. 😛

 

[DS426]

I don’t even think it’s a big deal for professional use. I’ve been running DP1.4a at 4K 240Hz with DSC for over two years and never notice any issues. I now have a DP2.1 display that can do 4K 240Hz without DSC (I think… maybe only 180 Hz without DSC?) and don’t see a difference.

DSC’s “visually lossless” compression does work well in practice.

I personally also felt DP 1.4a's DSC is subjectively visually lossless and sufficient for even serious photo, graphics, etc. work. I really shouldn't have said a "big deal" but a "bigger deal" relative to gaming, even if only slightly / only for certain use-cases.

 

[DS426]

The awakening might be brutal. We are already talking about 850$ for the 9070 XT (according to Jaytwocents). Those MSRP prices might not be seen anytime soon. You are all praising AMD right now but we might end-up with a card slightly more powerful than the 7900 XT starting at 850$. I would wait before saying Nvidia is cooked.
...

I think most agree that nVidia would be "cooked" *IF* supply was strong. A lot of praise appeared to be regarding the MSRP pricing given the performance of the cards as we had just recently seen rumors that 9070 XT pricing could be $700.

9070 XT is slightly less powerful than the 7900 XTX, not slightly more powerful than the 7900 XT; it's noticeably closer in perf to the XTX, especially when RT is at least considered. That's probably one way that Jay came to the $850 number as this would make sense when the 7900 XTX was selling for ~$900 (before these launch failures, anyways).

 

[JayGau]

I think most agree that nVidia would be "cooked" *IF* supply was strong. A lot of praise appeared to be regarding the MSRP pricing given the performance of the cards as we had just recently seen rumors that 9070 XT pricing could be $700.

9070 XT is slightly less powerful than the 7900 XTX, not slightly more powerful than the 7900 XT; it's noticeably closer in perf to the XTX, especially when RT is at least considered. That's probably one way that Jay came to the $850 number as this would make sense when the 7900 XTX was selling for ~$900 (before these launch failures, anyways).

A card with the 7900 XTX performance at 900$ is just a 7900 XTX with another name. The 7900 XTX was 1000$ at launch (for real, I got one a few weeks after launch at MSRP). Looking at the prices right now for the 9070 XTX they are pretty much all above 750$. I even saw one at 999.99$. And it's going to get even worse with the announcement that "only the first batch was meant to be at MSRP".

And I'm sorry but considering the fact that the 7900 XTX is already slightly more powerful than the 7900 XT, so my comment still stands.

 

[Manumarblanc]

https://twitter.com/x/status/1897668346298220588

View: https://x.com/AzorFrank/status/1897668346298220588

 

[-Fran-]

https://twitter.com/x/status/1897668346298220588

View: https://x.com/AzorFrank/status/1897668346298220588

Well, look at that. Looks like AMD is learning.

Thanks for sharing.

Regards.

 

[King_V]

Paper Launch. 90% of people paying less than $750 seems to have had their order cancelled. Gee THANKS AMD !!! THANKS FOR 3 MONTHS OF NONSTOP GASLIGHTING !!!

The scalpers, having run out of Nvidia cards to scalp, have decided to scalp the new AMD cards, and this is, somehow, AMD's fault?

Like, what? AMD set up the bots for the scalpers and all that?

 

[thestryker]

It would "work," but imaging still takes time and the images would be quickly outdated. I probably download at least 200GB of updates to games in my test suite every month. Sometimes more. So if I make an image today, I will still have 30 minutes or whatever of downloading updates once I put the image back in place.

And that's not accounting for the time it takes to actually reimage the drive — about 3TB of data, so even at ~1 GB/s that's nearly an hour. (That 1 GB/s figure is just an estimate based on SSD speeds and the imaging tools, as you rarely would get full write speeds of ~3 GB/s.)

Realistically, if I reimage between every GPU swap, I'm adding at least two hours to testing. And people already complain that "you didn't include the RTX 4070 Super because you're trying to hide things!" Yeah, right. It's not that I'm one person with one testbed doing all the work. It's that I'm trying to hide stuff! LOL

I want to preface this with my opinion that the way you do it now is fine.

If you had 3 or so identical SSDs ready to go with everything but video drivers installed this problem is entirely negated. Create an image after running updates when you know you're going to be testing. Then reimage as needed while running tests using one of the other drives. This should add a negligible amount of time to testing and if you didn't want to swap M.2 drives every time for fear of the shoddy connector M.2 to U.2 adapter/cable on the motherboard side and U.2 to M.2 enclosure on the drive side solves that one.

Again I think the way you do it now is fine, but it's not a huge stretch time wise to do effectively a clean install every test.

 

[JarredWaltonGPU]

I want to preface this with my opinion that the way you do it now is fine.

If you had 3 or so identical SSDs ready to go with everything but video drivers installed this problem is entirely negated. Create an image after running updates when you know you're going to be testing. Then reimage as needed while running tests using one of the other drives. This should add a negligible amount of time to testing and if you didn't want to swap M.2 drives every time for fear of the shoddy connector M.2 to U.2 adapter/cable on the motherboard side and U.2 to M.2 enclosure on the drive side solves that one.

Again I think the way you do it now is fine, but it's not a huge stretch time wise to do effectively a clean install every test.

FWIW, only three of the gaming tests in my suite are now even remotely automated — five if you count the two Flight Simualtors. And if you don't babysit the plane, it will crash and you need to drop back to the menu, requiring twice as much time per iteration. The pro/AI tests are less involved in terms of sitting in front of the screen and actively running them, but most of the games are fully manual tests.

Because "it's the only way to be sure." If I'm not looking at and running the tests, I don't really know what things feel like or if they're rendering the way I'd expect. Automation is super nice for making it easy, and super easy to make problems go unnoticed.

Unfortunately, I don't have multiple sets of identical hardware. That would be nice, but even then I can't test two games at the same time. I could do pro/AI on one and games on the other, perhaps, which is what I did the past two years. But I'm unified on Ryzen 9800X3D for the time being.

 

[Gururu]

Microcenter just dropped a ton of XTs for MSRP in MD.

 

[alceryes]

That ain't my problem, my micro center has around 40 in stock, it's that all of them are $250 or more above MSRP. At that point, it ain't super competitive anymore(I know it beats the 5070 ti, which is at the same price, sometimes but that ain't the point).

Was that after they sold out of their at MSRP stock? 'cause I think the NJ store eventually only had the not-worth-it high priced versions around 1-2pm or so. The MSRP stock was definitely the first to go.

 

[John Freiman]

SPECworkstation 4.0 -- while I appreciate that you did do some multimedia testing, I'm disappointed that you didn't disclose which Codecs you were transcoding from and to.
I assume the Spec test was H264 - because I think that's the default, but I, and I think most users who are interested in transcoding are going to be using a more modern codec like HEVC/x265 -- not the 22-year-old H264.
I hope to see a real comparison between the media engines on Nvidia, AMD and Intel in the coming weeks.
Plex has been supporting x265 transcodes since late last fall in alpha testing a couple/few months in the release channel.
I'd really like to know how RDNA 4 compares using Handbrake and FFmpeg with modern formats.
I want to upgrade my A380 and would like an efficient in codec quality, power efficiency and most importantly $$ efficiency!
Re: Plex, This year they are working on releasing an updated FFmpeg version going from v4 to v7... should be interesting.

 

[art.cabbie]

I got a 9070XT at Micro Center at 11AM. They asked which one do you want?
I said "The Cheapest One"
This is the only one left at MSRP $599.00 (Asrock Steel Legend)
me: "Sold!"
Then I had to wait 45 min in line to the cashiers to pay for it. The line for GPU went all the way to the back of the store.

Just for LOL's I looked at the MC website at 10PM.
Only card listed was $800, and Sold Out.
I am very happy with my purchase. It is 2X - 3X faster than my 3060 Ti. I can crank all the settings to Ultra quality with plenty of FPS.

 

[-Fran-]

May be relevant to the discussion?

View: https://www.youtube.com/watch?v=r-MYIjtSm_8


I'll link it in some other threads, so apologies if you see it repeated.

Regards.

 

[King_V]

I sometimes wonder if it works like this:

• Scalpers buy up all of one release
• Other scalpers buy up all of the next release
• Yet other scalpers buy up the higher priced stuff
• Yet OTHER other scalpers buy from the earlier scalpers, hoping to be able to make some profit over and above the original scalpers' prices

 

[stanjd]

Considering you can't buy a previous gen card for MSRP still, I doubt many people are getting these for $599. You can't get a 7800xt or a 4070 for that price Maybe the scalpers running bots can. They got all of the good deals anyway.

 

[JarredWaltonGPU]

Considering you can't buy a previous gen card for MSRP still, I doubt many people are getting these for $599. You can't get a 7800xt or a 4070 for that price Maybe the scalpers running bots can. They got all of the good deals anyway.

Apparently everyone needs a new PSU or mobo to go with their RX 9070 XT card. LOL



Also note that in most cases, those are $100 PSUs, and the cards are being listed as $729.99. That's for the Sapphire Pulse RX 9070 XT, which is most definitely an MSRP design. So the markup right now is "only" $130 if you get the combo. But you also need to return the whole combo if you have issues, which can be a pain as well if you want to return anything.

Things are much worse looking at PC Part Picker:

 

[KyaraM]

Funny how the benchmark suite can influence rankings. Techspot has the 9070 behind the 4070Ti in raster and the 9070XT behind the 4070Ti Super (at 1440p), here they are both faster. GN is different yet again, with suspiciously low results for the 5070 making me wonder if there was an issue with their card. Just shows that you really need to check different sites to compare, not just one, I guess...

AMD Radeon RX 9070 Review

AMD impressed us with the Radeon 9070 XT, but now we turn to its cheaper sibling, the RX 9070. Priced at $550, it competes with Nvidia's RTX...

www.techspot.com

In regards to availability, in Germany, Alternate got some stock of the 5070 and the 5070Ti right now (though the Ti model in question is pretty overpriced...), but none of the AMD cards, they are sold out. Both are comparable in price to their Nvidia counterparts, though the 5070 is currently cheaper than the 9070. This holds true when looking at comparison pages like Idealo.

 

[alceryes]

...Both are comparable in price to their Nvidia counterparts, though the 5070 is currently cheaper than the 9070. This holds true when looking at comparison pages like Idealo.

But that's just it - they shouldn't be comparable in price. AMD purposefully slotted the MSRP of the 9070 XT substantially below that of the 5070 Ti.

Unfortunately, AMD cannot control the actual street price. MSRP is a suggested price not a has-to-be price. It seems that brick-n-mortar stores are the only way to buy sought-after GPUs at MSRP now that the bots, scalpers, and even scalping online store (yes, they are doing it too) have found a new revenue stream in overpriced top end gaming GPUs.

 

[KyaraM]

But that's just it - they shouldn't be comparable in price. AMD purposefully slotted the MSRP of the 9070 XT substantially below that of the 5070 Ti.

Unfortunately, AMD cannot control the actual street price. MSRP is a suggested price not a has-to-be price. It seems that brick-n-mortar stores are the only way to buy sought-after GPUs at MSRP now that the bots, scalpers, and even scalping online store (yes, they are doing it too) have found a new revenue stream in overpriced top end gaming GPUs.

I'm not sure what you want to tell me here. All I did was outlining the pricing situation in one part of the globe. I didn't say anything about how anything should be or whatever, nor did I pass judgement on anything. and why shouldn't they be comparable in price anyways? Especially the 5070 and the 9070, which literally had the exact same MSRP of $550? The more amazing part is that the lowest listing of the 5070Ti is actually 50€ below MSRP, though it is, of course, not available.

 

[alceryes]

I'm not sure what you want to tell me here. All I did was outlining the pricing situation in one part of the globe. I didn't say anything about how anything should be or whatever, nor did I pass judgement on anything. and why shouldn't they be comparable in price anyways? Especially the 5070 and the 9070, which literally had the exact same MSRP of $550? The more amazing part is that the lowest listing of the 5070Ti is actually 50€ below MSRP, though it is, of course, not available.

Sorry, I'm just frustrated at these scalping stores taking advantage of GPU shortages and demand. Not necessarily directed at you, per se.

 

[thestryker]

Funny how the benchmark suite can influence rankings. Techspot has the 9070 behind the 4070Ti in raster and the 9070XT behind the 4070Ti Super (at 1440p), here they are both faster.

If you look at the individual game results in the HUB/Techspot review you'll see there are a couple of games they use that significantly disadvantage the 9070 cards. I noticed this in their 9070 XT review where there were some titles RDNA 4 performed much worse proportionally to RDNA 3 which skews their results (CS2 is a big one that was the same for every reviewer using it). While I'm always a big advocate for knowing the outliers when the game count is lower they throw off averages.

 

[JarredWaltonGPU]

If you look at the individual game results in the HUB/Techspot review you'll see there are a couple of games they use that significantly disadvantage the 9070 cards. I noticed this in their 9070 XT review where there were some titles RDNA 4 performed much worse proportionally to RDNA 3 which skews their results (CS2 is a big one that was the same for every reviewer using it). While I'm always a big advocate for knowing the outliers when the game count is lower they throw off averages.

Technically, I think it's important to note that I show the 5070 Ti as faster than the 9070 XT in all four of the aggregate rasterization charts. The 9070 meanwhile is faster than the 5070 in all four aggregate rasterization charts. But I guess KyaraM was looking at the 4070 Ti and 4070... which... why?

And by that, I mean why is that the main comparison? You can't really buy 4070 class cards at a reasonable price now, so while the comparison point is important on some level, it's not my primary comparison. New AMD vs old Nvidia? Just like I wouldn't compare new Nvidia vs old AMD as the main item. New AMD vs old AMD makes more sense, or new Nvidia vs old Nvidia.

Anyway, I know with brand-new, up-to-date test results (I'm not relying on numbers run sometime last year or even last month), the 9070 XT ends up quite a bit faster than a 4070 Ti Super. The 9070 ends up just barely faster than a 4070 Ti — 2% faster at 1080p medium and 1440p ultra, and 1% slower at 1080p ultra, but then 6% faster at 4K ultra. That makes more sense considering 12GB vs 16GB.

I don't have really any old and undemanding games in my current test suite. Flight Simulator 2020 definitely hits CPU limits on these cards, and Baldur's Gate 3 is pretty tame in GPU requirements as well. Just about everything else ends up fully GPU limited, and that will factor into the overall rankings.

 

[stuff and nonesense]

There are lies, damn lies and benchmarks.
All that can be done is to look for how your individual use case is covered across the benchmarks, not just TH, not just GN or HUB but across the spectrum of reviewers.

With that in mind, I lament the loss of written word sites. I prefer to not take notes but to flick between tabs on Firefox.