600 euros for a 16gb 60 class card.
I want a dozen please.

Meh, will be worse than the 5060TI and that's pretty bad.

AMD waits for the RTX 5060 before announcing their cards. They are also probably waiting to see if the RTX 5060 Ti will see low availability and inflated price latter. That's what I believe. And Nvidia said a simple "May" for the RTX 5060, without a specific date, to just troll AMD. Nvidia can wait until the end of May and sit there laughing at AMD that will be waiting for the official announcement of the RTX 5060.

I believe AMD will have to move faster this time. The waiting game payed off with 9070 series, but I doubt it will with the 9060 series. And hope their pricing to be lower, like $379 for the 16GB model and $299 for the 8GB model(and no more than $479 for the 9070 GRE). If they also come out sooner rather than latter with an RX 9050 that performs better than an RX 6600 at $199, while offering the advantages of RDNA4, that would be also a good product, considering market conditions.

john_AMD waits for the RTX 5060 before announcing their cards. They are also probably waiting to see if the RTX 5060 Ti will see low availability and inflated price latter. That's what I believe. And Nvidia said a simple "May" for the RTX 5060, without a specific date, to just troll AMD. Nvidia can wait until the end of May and sit there laughing at AMD that will be waiting for the official announcement of the RTX 5060.

I believe AMD will have to move faster this time. The waiting game payed off with 9070 series, but I doubt it will with the 9060 series. And hope their pricing to be lower, like $379 for the 16GB model and $299 for the 8GB model(and no more than $479 for the 9070 GRE). If they also come out sooner rather than latter with an RX 9050 that performs better than an RX 6600 at $199, while offering the advantages of RDNA4, that would be also a good product, considering market conditions.

i should laugh at people still talking about MSRP, but it's so sad i can't.
I'm sure this time will be different :roll:

Perhaps when they release the 9060 we get a bette driver. I'm still forced to use 24.5.1 due problems with Delta Force.

Bomby569i should laugh at people still talking about MSRP, but it's so sad i can't.
I'm sure this time will be different :roll:

Considering MSRP is a starting point, it remains important even if it is fake, so yes, people will keep referring to it.

john_Considering MSRP is a starting point, it remains important even if it is fake, so yes, people will keep referring to it.

starting point would imply it was something real

I have very low expectation for the RX 9060 series because,
1. It is still on 128bit memory bus,
2. GDDR6 = No meaningful improvement to memory bandwidth to feed the memory bandwidth starved GPU, whereas the RTX 5060 Ti is using GDDR7,
3. Mostly no increase in the specs like number of CUs, ROPs, etc. Clockspeed should increase, though I don't think it is very substantial.

I believe the RX 7600 series did not suffer from latency issues like the bigger Navi 31 and 32 since it is a monolithic chip, which is also the case here. Therefore, other than improvements in RT and AI performance, I feel the RX 9060 will underperform the RTX 5060 Ti. So unless AMD price this very aggressively, I feel this will not be a popular card.

At this point we just need to get as many graphics cards out there as possible and hope they start dampening demand and bringing down prices both new and prior gen used.

I'm less excited about the 9060 series than I was for the 5060Ti.

A significant problem Nvidia had with the low-end 4060 series was bandwidth, solved by changing to GDDR7 with the 5060 series. 15-18% more performance than the 4060Ti isn't exactly impressive, but it's by far the most significant 40-vs-50-series improvement so far, likely because the 4060Ti was so utterly hamstrung by its 128-bit GDDR6 bus.

The 9060 series is still GDDR6 and that puts it at a non-trivial disadvantage. Both the 7700XT and 7900GRE showed that bandwidth was one of their more significant bottlenecks since memory overclocks ONLY yielded significant performance gains on those particular models - certainly more significant than memory overclocking on the 7800XT which had less shader power per GB/s of bandwidth by default.

7800XT @ 624GB/s for 3840 shaders = 166MB/s per shader
vs:
7700XT @ 432GB/s for 3456 shaders = 128MB/s per shader (23% less bandwidth)
7900GRE @ 576GB/s for 5120 shaders = 115MB/s per shader (31% less bandwidth)

I can't really compare bandwidth per shader of the 9060 series against RDNA3 since the IPC per shader core is different and the clocks are far higher, but that just hurts its relative memory bandwidth even harder since that is still stuck on 20Gbps GDDR6 regardless of the IPC and clock gains.

In terms of expected compute power vs bandwidth, the 9060 series is worse off than either of those two bandwidth-limited RDNA3 cards. If the 9060 XT is half the raw compute of the 9070XT then it's going to really struggle for bandwidth above 1080p, if it doesn't also cause issues at 1080p!

Chrispy_I'm less excited about the 9060 series than I was for the 5060Ti.

A significant problem Nvidia had with the low-end 4060 series was bandwidth, solved by changing to GDDR7 with the 5060 series. 15-18% more performance than the 4060Ti isn't exactly impressive, but it's by far the most significant 40-vs-50-series improvement so far, likely because the 4060Ti was so utterly hamstrung by its 128-bit GDDR6 bus.

The 9060 series is still GDDR6 and that puts it at a non-trivial disadvantage. Both the 7700XT and 7900GRE showed that bandwidth was one of their more significant bottlenecks since memory overclocks ONLY yielded significant performance gains on those particular models - certainly more significant than memory overclocking on the 7800XT which had less shader power per GB/s of bandwidth by default.

7800XT @ 624GB/s for 3840 shaders = 166MB/s per shader
vs:
7700XT @ 432GB/s for 3456 shaders = 128MB/s per shader (23% less bandwidth)
7900GRE @ 576GB/s for 5120 shaders = 115MB/s per shader (31% less bandwidth)

I can't really compare bandwidth per shader of the 9060 series against RDNA3 since the IPC per shader core is different and the clocks are far higher, but that just hurts its relative memory bandwidth even harder since that is still stuck on 20Gbps GDDR6 regardless of the IPC and clock gains.

In terms of expected compute power vs bandwidth, the 9060 series is worse off than either of those two bandwidth-limited RDNA3 cards. If the 9060 XT is half the raw compute of the 9070XT then it's going to really struggle for bandwidth above 1080p, if it doesn't also cause issues at 1080p!

Radeon GPUs are using Infinity cache to make up the "losses" in memory bandwidth. Also RX 9070 XT is also using GDDR6 and it's doing just fine.

@Chrispy_

I've actually made a post about that here.
Again the ratios are the same, I'm not saying that there isn't some bottleneck, although in order to prove that it is and how much it is a 9070 XT with just OCed memory should scale significantly up to a point, a point where there is no longer a bottleneck and thus further increasing memory frequency doesn't yield additional gains.
So that cache magic definitely works, well... until it doesn't, but most of the time it apparently works, otherwise the gaming performance would be affected.
In other workloads, the software support isn't complete so there are still tests that should be done, but Blackwell also has support issues.
The 9060 XT will probably be a couple of percent behind the 5060 Ti in the relative performance chart. The price will make or break it, AMD could get greedy for all we know.

RejZoRRadeon GPUs are using Infinity cache to make up the "losses" in memory bandwidth. Also RX 9070 XT is also using GDDR6 and it's doing just fine.

Those GPUs I mentioned (I've used all three of those GPUs and still own one of them) all had infinity cache, so it's like-for-like and there's no reason to bring up GPU cache unless you're comparing to older Nvidia cards (Nvidia increase their cache by around 8x for the 40-series).

The 9070XT is doing fine because it has twice as much bandwidth as the 9060-series and twice as much cache. Memory bandwidth requirements largely scale with render resolution and graphics settings. Given that we're talking about GPUs in the $400-800 range, I don't think it's unreasonable to expect the same 1440p resolution here, since decent 1440p high-refresh monitors start at about $180 and if you can afford even a $400 GPU that will last you 3-5 years for decent AAA gaming, you can definitely afford a $180 monitor that will last you a decade and it's a false economy to spend so much money on a GPU and water down your experience on an ancient low-res display.

HUB roundup including best budget gaming monitors at the 4:46 mark:

All I say is wait and see. RX 9070 XT was a surprise and non-XT beats RTX 5070 in almost everything. Also the pricing of RTX 5060 can be whatever but probably anything but MSRP seeing how totally screwed up this part of the market has been for years now.

Based on the existing numbers, where would you put this card in terms of strength? Between 7700XT and 4060ti? But at what price? I think it would be viable for $350 at the most

@Grebber01

I talked about that here.
I am convinced that the 9060 XT will beat the 4060 Ti, therefore the expected performance level should be at least similar to 3070 Ti / 7700 XT which are marginally lower than 5060 Ti.
Depending on how the RDNA4 architecture scales performance should be close to the 5060 Ti, I still consider that it will not beat the 5060 TI even if it will be very close to it in raster.
www.techpowerup.com/review/palit-geforce-rtx-5060-ti-infinity-3-16-gb/33.html
Price-wise I would say that $350 for the 16GB version is very reasonable as an upper limit (compared to the competition) and if somehow AMD is greedy and puts a higher MSRP than that I strongly feel that they should be booed off the stage.

We can already see that the market did not adjust the higher-than-normal-MSRP-influenced-street-price of the 9070 downwards, instead it pushed the 9070 XT upwards closer to the street price of the 5070 Ti than the ratio between the two cards' MSRPs suggests. Basically the retailers look at the performance numbers, if they're close they'll adjust the street prices according to that as well and not just by the MSRP.

If we are looking at 3.200MHz / 64ROPs / 128T.U. / 128bit bus with 20Gbps GDDR6, and the RX 9060 XT specs are identical between 8GB and 16GB versions then it will be something like the below in QHD.
Depending on the game selection of the testbed the 16GB RX 9060 XT can have higher difference with the 8GB version if there are many memory demanding games, here I just assumed a minimum 1%, it should be even higher.
Even at $300 MSRP the 8GB version of RX 9600 XT will have a problem in relation with RTX 5060 and regarding the 16GB version, to make an impact it must be $350 (if in Nvidia's MSRP the difference between 16GB & 8GB GDDR7 is $50, AMD should also set $50 for GDDR6, it will look bad otherwise, it's better to offer the same performance/price difference for RTX 5060Ti 16GB vs RX 9060 XT 16GB as what we have in RTX 5070 vs RX 9070 XT, than to lose another opportunity again to gather positive reviews.
Of course it will all depend on the actual street prices, but the % differences should be according to the below table:
QHD performance chart:

RTX 5060Ti 16GB	111	$430
RX 9060 XT 16GB	101	$350-$370
RX 9060 XT 8GB	100	$300
RTX 5060 8GB	95	$300

AMD is again with the knife and cheese in hand (an expression in Brazil for something like 'Have all the aces'). But again, AMD cuts its own hand with the knife and drops the cheese on the floor.