Recent content by ArkeoTP

So, I guess that puts the "AMD is getting exclusive access to 1GHz chips" rumour to rest.

Some benchmarks from reddit: From:

But will the miners buy it? Given the almost nonexistent resale value of these things. Of course, they may buy it regardless because there quite literally isn't anything else to buy but will it ever reach that point?

But you may want to turn off TAA more often than not with a mGPU setup as temporal techniques aren't AFR friendly and can cause problems with scaling and frame pacing. But yeah, you always need more AA* Increasing spatial resolution helps everything somewhat but it's not the solution to end...

So, you're telling me that Intel is actually helping AMD's performance here? Dear Lord, what the heck is up with AMD's OpenCL driver?

Unless it was changed in an update, async was only active when using either no AA or TSSAA. Any of the other AA options disable async.

PCPer's Vega FE Crossfire benchmarks are out. https://www.pcper.com/reviews/Graphics-Cards/AMD-Radeon-Vega-Frontier-Edition-CrossFire-Testing It doesn't look good albeit I'm not sure I should be surprised as a former Crossfire user.

So, I was wondering. While Afterburner and other applications allow me to roughly check what graphics API a game is running, it doesn't give more precise details. Is it possible to check which specific API version (feature level?) a game is running (and used extensions in the case of OpenGL)?

Yeah, there's a dedicated PISS (Path Integral Solution Solver) that modulates quantum effects to encode clocking and timing information in sub-cycle timespans which allows it to overclock for brief amounts of time (< one cycle). Do note that you are also able to overclock the PISS, but beware...

PCPer updated their "doe" size measurement to 512mm^2 There's also this tweet by Raja Koduri, which implies a die size of 484mm^2 (22*22) So ummm... Yeah...

I really hope not but it's possible they messed up the IMC again and the card's effective memory bandwidth is far below the theoretical bandwidth. I don't think it's enough to explain the regression though.

Ok, second take. I revised the borders, which puts it around 26.25*19.75 = 518mm^2 It isn't a pixel perfect measurement but it should be more or less accurate. Maybe PCPer measured it wrong? Or maybe I measured it wrong?

I placed the photo in Krita, resized the card to line up with the ruler at 267mm, and the chip looks to be about 28mm by 20mm putting it at 560mm^2. PCPer's measurement appears to be accurate.

More detailed PCB overview: From: https://www.pcper.com/news/Graphics-Cards/Radeon-Vega-Frontier-Edition-GPU-and-PCB-Exposed

It's going to be one of those cards again where reducing voltages actually improves performance. Depending on your luck of course.