ASUS Radeon RX 9060 XT Prime OC 16 GB

[Hecate91]

Ignoring a few outliers and the 9060XT is competitive with the 5060Ti 16GB, thats not cherry picking.

As for CS2, 200fps is still more than enough for a fast paced FPS title, anyone in e-sports playing for a team are getting a gaming PC from a sponsor anyway.

 

[GodisanAtheist]

Well it performed almost exactly like we thought it would.

If you think the performance is anything other than we expected, then it's your expectations that we're wrong.

It's a 7600xt with a 40% RDNA4 uplift.

 

[SL2]

The N44 is basically half of the N48 in almost every aspect, except it doesn’t have half the number of transistors. Odd.

Is everything cut in half? Like display engine and PCIE controller, are you sure?

 

[medi01]

30% faster than 7600XT for 10% more $.
Pretty nice even by 2020 metrics, and simply decent by today's "nGridia gone mad" takes.

As for "9070 should not be... that much faster"... what?
For nearly 100% more monez you get what, 50% more perf.
And that is "too much perf" gap?

 

[AusWolf]

A tad slower than the 5060 Ti, but a fair bit cheaper as well. Just as expected.

 

[suraswami]

Glad I didn't buy the 5060 non-ti 8GB crap. I will wait for initial launch craze to settle down and hopefully MC keeps the price honest.

 

[JustBenching]

30% faster than 7600XT for 10% more $.
Pretty nice even by 2020 metrics, and simply decent by today's "nGridia gone mad" takes.

As for "9070 should not be... that much faster"... what?
For nearly 100% more monez you get what, 50% more perf.
And that is "too much perf" gap?

Ngreedia gone mad huh? The card you are prasing is 3% faster than a 5060 at 1080p while it costs 18% more. Ngreedia, yeap, you just can't help it.

 

[dgianstefani]

Ignoring a few outliers and the 9060XT is competitive with the 5060Ti 16GB, thats not cherry picking.

As for CS2, 200fps is still more than enough for a fast paced FPS title, anyone in e-sports playing for a team are getting a gaming PC from a sponsor anyway.

It's the definition of cherry picking. The point of TPU reviews is consistency as much as possible, for control all else is equal, besides perhaps launch drivers for new cards, the experiment is the new card being tested. You're suggesting breaking an entire test methodology used in every other review, so that one card with bad results (something people would want to know about) looks better in graphs. For what reason? I can certainly guess. If it is a driver issue then it might be resolved with next test suite using newer graphics drivers, if not, then AMD cards not working well with these games is also relevant information to consumers making buying decisions.

Debating about 200 vs 300 FPS is irrelevant, esports gamers (like the ones who play CS2 and buy entry level discrete cards) would disagree on your analysis of importance, and even if it was relevant, 50% extra FPS is not something to ignore.

 

[lexluthermiester]

Ngreedia ... Ngreedia ...

Really? When is this nonsense going to stop?

 

[SL2]

Ng

Conveniently leave out the difference in VRAM, that's high praise!

 

[Denver]

It's the definition of cherry picking. The point of TPU reviews is consistency as much as possible, for control all else is equal, besides perhaps launch drivers for new cards, the experiment is the new card being tested. You're suggesting breaking an entire test methodology used in every other review, so that one card with bad results (something people would want to know about) looks better in graphs. For what reason? I can certainly guess. If it is a driver issue then it might be resolved with next test suite using newer graphics drivers, if not, then AMD cards not working well with these games is also relevant information to consumers making buying decisions.

Debating about 200 vs 300 FPS is irrelevant, esports gamers (like the ones who play CS2 and buy entry level discrete cards) would disagree on your analysis of importance, and even if it was relevant, 50% extra FPS is not something to ignore.

Lightweight games like these should be excluded, as they run at hundreds of FPS on virtually any modern hardware from the past five years, making them irrelevant for meaningful GPU testing. They also skew the average, making it less representative of performance in recent AAA titles.


If they're not removed from the review entirely, they should at least be excluded from the average framerate calculation.

 

[oxrufiioxo]

Ngreedia gone mad huh?

To be fair both have gone AI.

I wouldn't classify that as mad though they are just following the money.

I have a bad feeling the next generation is going to be even worse than this one minus the 6090 becuase both companies are so focused on AI

I do find it comical that yet again people are arguing over two meh AF gpus.

Yay we got a 4060ti 16G like product almost 2 years later for less money allegedly.

The biggest dissappointment is that 60 class cards from 2025 haven't beat Nvidia 70 class cards by a noticeable margin from 2020 smh #progess

 

[lexluthermiester]

Lightweight games like these should be excluded

Not when they are very popular and show valid performance numbers.

 

[Hecate91]

Ngreedia gone mad huh? The card you are prasing is 3% faster than a 5060 at 1080p while it costs 18% more. Ngreedia, yeap, you just can't help it.

The card you're bashing on has better price/performance at 1080P than the 5060ti though,lol.

It's the definition of cherry picking. The point of TPU reviews is consistency as much as possible, for control all else is equal, besides perhaps launch drivers for new cards, the experiment is the new card being tested. You're suggesting breaking an entire test methodology used in every other review, so that one card with bad results (something people would want to know about) looks better in graphs. For what reason? I can certainly guess. If it is a driver issue then it might be resolved with next test suite using newer graphics drivers, if not, then AMD cards not working well with these games is also relevant information to consumers making buying decisions.

Debating about 200 vs 300 FPS is irrelevant, esports gamers (like the ones who play CS2 and buy entry level discrete cards) would disagree on your analysis of importance, and even if it was relevant, 50% extra FPS is not something to ignore.

Excluding two games isn't cherry picking and it doesn't break any testing when you can see every other game on the list performs well, even Nvidia sponsored AW2.
I couldn't care less about esports BS, the elitist "but muh 500 bazillion FPS" nonsense is a major reason why I couldn't give a sh*t, but getting 200fps vs 300fps for a casual player won't make any difference at all at 1080p. Like I said anyone making money off the game isn't going to be using a 60 class card.
Also saying a few games perform poorly is going to steer customers away is disingenuous especially since you're ignoring how bad Nvidia drivers have been lately for 40 and 50 series cards. The 9060XT is as expected, and better value if these cards are close to MSRP. I'm not surprised the Ngreedia fans are looking for anything to criticize it with though.

 

[JustBenching]

Conveniently leave out the difference in VRAM, that's high praise!

I thought 8gb of vram cost 10$. You are upcharged for vram while getting the same performance. Lovely

The card you're bashing on has better price/performance at 1080P than the 5060ti though,lol.
View attachment 402471

I have astigmatism on the left eye and the text you posted has really tiny letters, can you help me out here, what does it say up there at the top? What card is it?

 

[Chrispy_]

The N44 is basically half of the N48 in almost every aspect, except it doesn’t have half the number of transistors. Odd.

The transistor budget has to include stuff that doesn't scale with the compute cores or memory controller.

Stuff like fixed function hardware, HDCP modules, display controllers, media engine, clock/power control logic, chip management, and other SoC stuff.

I'm making a guess here, but I believe
EDIT: I found a block diagram, so the stuff I've highlighted is mostly copy+pasted onto smaller chips. This 'uncore' cannot be halved like the shader engines and memory bus can.

 

[Hecate91]

I thought 8gb of vram cost 10$. You are upcharged for vram while getting the same performance. Lovely

You're getting the same amount of VRAM for $80 less. It's around 30-35% faster than the 7600XT for 10% more money, which is a better price/performance value than what the jacket man provided with the 5060Ti.

I have astigmatism on the left eye and the text you posted has really tiny letters, can you help me out here, what does it say up there at the top? What card is it?

Did you miss what I said too?
The B580 is also higher on the chart but those aren't selling anything close to MSRP either.

 

[wNotyarD]

I have astigmatism on the left eye and the text you posted has really tiny letters, can you help me out here, what does it say up there at the top? What card is it?

Certainly not the 5060 Ti.

 

[JustBenching]

You're getting the same amount of VRAM for $80 less. It's around 30-35% faster than the 7600XT for 10% more money, which is a better price/performance value than what the jacket man provided with the 5060Ti.

Did you miss what I said too?
The B580 is also higher on the chart but those aren't selling anything close to MSRP either.

Certainly not the 5060 Ti.

So nobody is gonna help an old man out. Okay

LTT says he wasn't sent an 8gb version of this. Is it true?

 

[lexluthermiester]

LTT says he wasn't sent an 8gb version of this. Is it true?

Who cares? I mean, do we really care about LTT?

 

[Neo_Morpheus]

Lightweight games like these should be excluded, as they run at hundreds of FPS on virtually any modern hardware from the past five years, making them irrelevant for meaningful GPU testing. They also skew the average, making it less representative of performance in recent AAA titles.


If they're not removed from the review entirely, they should at least be excluded from the average framerate calculation.

Agreed.

Does it really matters to a player in a tournament that his hardware is pushing over 200 or more FPS?

But as stated, it does wonders for the averages.

 

[JustBenching]

Who cares? I mean, do we really care about LTT?

Don't you want to see reviews of it?

 

[lexluthermiester]

Don't you want to see reviews of it?

Not from LTT. Or HUB, J2C or GN for that matter. Objectivity and impartiality matters. They can not be counted on for such things.

 

[Squared]

A 6-series card in 2025 with an x16 PCIe interface? Fabulous. It also has the same max power draw and size as my current graphics card so it'll run perfectly in my computer and use the full bandwidth of my PCIe 3.0 x16 slot. The last comparable graphics cards were the Arc A770, the RTX 3060 Ti, and the Radeon 5600 XT I have. I might buy one tomorrow.

But wait, 3 display outputs? I use 3: 2 for monitors and 1 for my TV. I can make that work. But either I have to get a DisplayPort model today and a display adapter, or an HDMI model today and a display adapter later when I get a new second monitor. I'd rather pay more and get 2 DisplayPorts and 2 HDMI ports. And I guess I'll never try a VR headset. I don't know about this. Why don't any models have a fourth port?!?

 

[Jism]

Sounds like a 6700XT is better in a few regards; meaning 1440p and maxed out stuff.

still good 1080p card tho for the price and value you get. Can't complain.