5700XT 50th edition

[alxlwson]

Specs on this look pretty good! The tech demos looked great. $499 only DTC from AMD seems like a pretty good deal. Can't wait to see some performance numbers.

 

[THUMPer]

I'm hoping partner cards get close to those speeds.

 

[Dan_D]

Honestly, the $50 increase seems like allot for what little it seems you get. Then again, it depends on what speeds the regular versions can attain and how you feel about the gold accents. Also, it depends on what difference there is in actual performance between the two.

 

[SeymourGore]

Yar, doesn't seem like a huge increase (performance wise) but I'm really glad that they bumped it up a little compared to the 50th 2700X.

 

[Dan_D]

Fair enough. It is nice to see, but I think the card itself tells us something far more interesting. I'm betting that high end Navi GPU's won't be much of an overclocker. If the clocks of the 50th anniversary card were so easily attainable, I think we would see those clocks be the standard operating frequencies of the RX 5700XT so that it would put even more distance between it and the RTX 2070 and even become more competitive with the RTX 2080.

Then again, it could be that those clocks are easily reached, but the power consumption and cooling requirements go into the stratosphere. Time will tell.

 

[DejaWiz]

This might be just the GPU at a price point I've been holding off on to upgrade my 980Ti. Fingers crossed!

 

[THUMPer]

Rumor has it AMD locked BIOS (no surprise) but also removed power play tables.

 

[Cranky 1970]

I still think it's a tad bit expensive for what it is. I hope it's a screamer though. I think i will still be rocking my Vega 64 with a full cover WB for as while though.

 

[noko]

Have 5700 XT 50th AV installed. With more aggressive fan curve and +25% power limit, it maintains clock speed in the high 19xxmhz GPU clock speed and sometimes goes over 2000mhz. Did not touch the Wattman Frequency or Voltage yet. This is exceeding rated Boost frequency but then I am allowing it to have more juice and air flow from stock. The stock fan curve really sucks for keeping the GPU cool and does limit the performance. Spinning up the fan makes a big difference.

Will most likely do an experiment with A/C cooling the whole case and drop the ambient temperature down. Depending upon how aggressive I want to be I could push for near to below freezing temperatures (just hard on fans and platters when the temperature gets low).

Performance is definitely faster than a Vega 64 LC that was in the machine.

 

[noko]

Did some OCing with the 5700 XT 50th AV, captured it here. I normally use FPS monitor to capture the parameters but that program just crashes with the 5700 XT so I had to do something else. Anyways a custom fan curve, Frequency OC to around 2100mhz, +25% power limit and it has no issue maintaining over 2000mhz. This is just with the stock cooler, air cooling and no modifications to the card.

 

[tungt88]

Did some OCing with the 5700 XT 50th AV, captured it here. I normally use FPS monitor to capture the parameters but that program just crashes with the 5700 XT so I had to do something else. Anyways a custom fan curve, Frequency OC to around 2100mhz, +25% power limit and it has no issue maintaining over 2000mhz. This is just with the stock cooler, air cooling and no modifications to the card.

That leads me to think that the upcoming custom 5700XTs from AIBs should be pretty good -- keeping my fingers crossed (wondering about 4K performance, since I play @ 4K).

This also gives grounds for a bit of cautious optimism: 5700XT

 

[noko]

Been gaming using pcie 4 on an ASUS X370 Crosshair 6 Hero motherboard. No issues, crashes etc. So far it has just been a very solid card. 3DMark PCIe Express test for PCIe 4 is 25.19GB/S:

I scored 0 in PCI Express

AMD Ryzen 9 3900X, AMD Radeon RX 5700 XT x 1, 32768 MB, 64-bit Windows 10}

www.3dmark.com

I did some test comparing PCIe 4 to PCIe 3 - gaming test -> no difference
Tried also Game + CPU game capture with Microsoft Game Capture -> No difference
Tried also Game + Bendicam CPU capture -> no difference
Used Strange Brigade DX 12 benchmark which gives very consistent results from one run to the next. Video capturing did reduce frame rates a little but in all capture methods, GPU, CPU, CPU - PCIe 4 or PCIe 3, the FPS result was exactly the same, not 1 FPS different! Which kinda threw me for a loop. Not usual when you see zero variation on game benches even after rebooting, changing from PCIe 3 and 4. Yes confirmed by 3DMark PCIe test.

Testing PCIe at 8x for PCIe 4 and PCIe 3 is the next logical step if one is going to use the second X16 slot for something like a SSD or even a second GPU.