AMD 7600 reviews

[BFG10K]

I think the product is below average/subpar but not terrible. The $269 price-point does help a little, even with the 8GB VRAM.

 

[Heartbreaker]

Review are nearly as bad as 4060...

GN review:

HWUB

 

[jpiniero]

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT / Die Taktraten unter Last

www.computerbase.de

7600 loves Callisto and COD MW 2... 20% faster. But most are barely faster than the 6650 XT. F1 22 it was even slower by a tad.

 

[coercitiv]

I think GN nailed it: AMD is a mess.

 

[Heartbreaker]

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT / Die Taktraten unter Last

www.computerbase.de

7600 loves Callisto and COD MW 2... 20% faster. But most are barely faster than the 6650 XT. F1 22 it was even slower by a tad.

Yes, it trades blows with the RX 6650. Which really points out the lack of performance improvement from RDNA 3, since these each have 32 CUs.

 

[biostud]

For some reason I don't regret buying a 6800XT in january.

 

[fralexandr]

Why when 6700 10gb are $280...
At least it's not like the 4060 series launch where some key stats actually make it look like it should be a 4050...

 

[In2Photos]

For some reason I don't regret buying a 6800XT in january.

Me either! If anything I'm really glad I didn't wait any longer. This generation of cards is CRAP!!!

 

[insertcarehere]

Can we finally stop pretending that RDNA3 architecture is not a dud? What are we seeing here that can't be replicated with an optimized 6650xt on 6nm?

 

[Heartbreaker]

Can we finally stop pretending that RDNA3 architecture is not a dud? What are we seeing here that can't be replicated with an optimized 6650xt on 6nm?

Yep. Its already near identical to 6650. 6650 is even ahead in a couple of games. RDNA3 mainly seems to improve RT over RDNA 2.

 

[AtenRa]

Can we finally stop pretending that RDNA3 architecture is not a dud? What are we seeing here that can't be replicated with an optimized 6650xt on 6nm?

I can only thing of RT performance, especially in heavy RT workloads like CP77, otherwise this is a meh, simple no perf increse directly from mArchitecture

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT / Die Taktraten unter Last

www-computerbase-de.translate.goog

 

[Timorous]

Can we finally stop pretending that RDNA3 architecture is not a dud? What are we seeing here that can't be replicated with an optimized 6650xt on 6nm?

AV1, DP 2.0

Per TPU numbers it is around 4% faster than the 6650XT in 1080p and has 98.5% of the clockspeed which is an IPC gain of 5.6%. Not great really.

 

[PJVol]

Anyone can link to the RX 6700 10gb review? Can't find any

 

[Timorous]

Anyone can link to the RX 6700 10gb review? Can't find any

AMD Radeon RX 7600 Review: Incremental Upgrades

A sideways step from the RX 6650 XT.

www.tomshardware.com

That has the 6700 10GB in the charts.

 

[insertcarehere]

I can only thing of RT performance, especially in heavy RT workloads like CP77, otherwise this is a meh, simple no perf increse directly from mArchitecture

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT

AMD Radeon RX 7600 im Test: Taktraten, Benchmarks in Full HD mit und ohne RT / Die Taktraten unter Last

www-computerbase-de.translate.goog

It is much better than 6650XT that's for sure... But also given how in this benchmark it's still decently slower than the 6700xt in similar proportion to the raster difference I am more inclined to believe something weird happened with the 6650XT there more than that there are dramatic improvements with the 7600.

 

[IEC]

Could have bought a 6600/XT for less and had that level of performance for like a year+.

Not moving the needle any even though it's the least badly priced current gen GPU. Which isn't saying much because they're all badly priced for the current market conditions.

Buy any used 12GB+ card for gaming. If you're paying for 8GB vRAM in 2023 it better be for a deeply discounted used card knowing it's relegated to entry level.

 

[In2Photos]

Pretty much...

https://twitter.com/x/status/1661361196786913280

 

[gdansk]

Can we finally stop pretending that RDNA3 architecture is not a dud? What are we seeing here that can't be replicated with an optimized 6650xt on 6nm?

N33 is on 6nm too. Never had high hopes for it. It pretty much is a N23 refresh even in conception. But as you say if RDNA3 wasn't a dud maybe they could have had 10% better performance instead of near equality.

 

[AtenRa]

In a parallel universe this card would have been produced at 5nm, be called RX 7500XT and priced $149

 

[Tup3x]

Another horrible product.... Not quite as bad as 4060 Ti but pretty bad still.

 

[VirtualLarry]

Another horrible product.... Not quite as bad as 4060 Ti but pretty bad still.

I don't think it's THAT bad of a product, but it doesn't seem to move the lower-end price/performance needle much if any, due to lack of generational improvement over it's predecessors.

 

[KompuKare]

Can we finally stop pretending that RDNA3 architecture is not a dud? What are we seeing here that can't be replicated with an optimized 6650xt on 6nm?

Maybe because aside from the RDNA3 changes, this almost is exactly that: a port of the Navi23 to 6nm. IF - and it is a very big IF - AMD were interested in a high volume part, then this part could be it. Yes, 8GB in 2023 isn't great but if they were to reduce the MSRP closer to $200 it could sell really well.

I still think that the go chiplet crazy strategy was far too risky. A 6nm monolith Navi32 able to perform closer to 6800 with 12GB or better still 16GB at maybe $350 would have sold very well IMO.

 

[VirtualLarry]

Yes, 8GB in 2023 isn't great but if they were to reduce the MSRP closer to $200 it could sell really well.

$200 would be a much better price. As it is, the ARC A750 8GB for $199 might just eat AMD's low-end lunch.

 

[blckgrffn]

As Tom's points out, when it comes to new cards, the performance to dollar scaling is nearly the exact same at the 4060 Ti, so they got the same 3.5 star grade.

That's damning with faint praise, and at the end they said they simply don't recommend buying either, and if you want the 7600 at least wait for the 4060 vanilla because that will likely cut its price down.

If this migrates to $199 by the end of summer and they phase out the 6600, then that's "fine". Definitely not exciting, especially given how much more power its using than the 6600 to produce its extra performance. It's nearly 50% more power usage compared to my bench 6600s which according to the AMD overlay top out at 100W out of the box.

 

[insertcarehere]

Maybe because aside from the RDNA3 changes, this almost is exactly that: a port of the Navi23 to 6nm. IF - and it is a very big IF - AMD were interested in a high volume part, then this part could be it. Yes, 8GB in 2023 isn't great but if they were to reduce the MSRP closer to $200 it could sell really well.

I still think that the go chiplet crazy strategy was far too risky. A 6nm monolith Navi32 able to perform closer to 6800 with 12GB or better still 16GB at maybe $350 would have sold very well IMO.

Well, the RDNA3 architectural changes (namely dual FP32 and higher clocks) were supposed to be substantial by themselves.

As it is now dual FP32 doesn't seem to do anything and without a significant process advantage it doesn't even seem to clock much higher on desktop than RDNA2 can.