Recent content by mczak

To be fair, it's a RX 6400 (not XT). FWIW, I think this could be a far better product than the RX 6500 XT. Navi 24 is cut down to the max, and then the 6500 XT goes in the opposite direction (with super high clock speeds, both on core and memory, sacrificing efficiency, increasing costs, with...

I think in theory it would be allowed to change all the tyres back to the old ones and send the driver back that way (assuming the wheel didn't get damaged otherwise). But due to the requirement of having to use 2 different compounds (not that it would be possible to use the same soft tyres for...

Yes and no. Verbally yes it was claimed it's faster. But on nvidia's own presentation slide the 3070 was actually "only" exactly as fast as the 2080ti. My interpretation would be that you have to include quite some titles which use rtx to get the same average performance as the 2080ti...

No idea about Kvyat, but for Max I'd guess that unforseen technical issues would be exactly such an exception listed "under paragraph A. 2. a) to g) of the Technical Directive 011-17" as quoted in the ruling.

Not saying the slide is wrong, just does not quite tell the full truth perhaps. From what I can tell from the open source drivers, CDNA really is GCN (Vega). Yes there might be some tweaks underneath here and there, but it seems like a stretch to claim this is really a different architecture. Of...

The conclusion makes sense to me too. We know that CDNA in the form of Arcturus is basically GCN (Vega - though I guess the G in the name would be a bit inappropriate here...) with some graphic bits stripped off. So CDNA2 could really be anything, and imho it makes a whole lot of sense if this...

I agree the 1750Mhz gpu clock are a paper number, nevertheless the actual average clock under load could easily be more than 60% higher than on Vega 10 (IIRC that number was more like ~800Mhz). A more efficient cpu (and potentially better power distribution between cpu and gpu) could also help...

Actually AMD changed this to 150W on their website: https://www.amd.com/en/products/graphics/amd-radeon-rx-5500 I don't disagree though 110W would make a lot more sense...

Variable rate shading comes to mind, as well as task/mesh shaders. There's a bunch of other (smaller) gl/vulkan extensions listed for turing too.

The separate int/fp units are there in volta as well. There is very little difference between turing and volta as far as the compute cores are concerned. The tensor functionality improvement is more of a tweak. Although there's more changes on the 3d graphics side. Calling it an evolution though...

Not sure why you think the penalty is too small? This is just the penalty you get when driving into pit lane from the wrong side of a pollard. Where I watched it they even called it out on TV that this was not correct and they are expecting a 5s penalty. Besides, I'm actually not sure he...

Don't insult my smartphone (SD 650 powered)! The frequency though seems to be quite low, the SD 650 already reached 1.8Ghz (but not sure it's the same 28nm tech). Although I guess the SD 650 has much better power / clock management (it also only has 2 A72 cores apart from the 4 A53 ones, but the...

Can't say I blame Ricciardo much for the first incident, though certainly if Vettel got a penalty for his incident last race this one has to be a penalty as well. Though I'd say the second one was a rather stupid move, it was obvious that he could only overtake there on the right side by going...

How is it decided whether to use wave32 or wave64 mode? I mean, the compiler could know the additional stalls with wave32 mode, but it seems like there would be a lot more factors which would decide what is optimal (and for things like potentially divergent branches that might be factors the...

The 16kB L1 of Vega 10 are per CU, this one appears to be a per-chip number (though is that really kb and not kB?). Doesn't fit the number of CUs though, maybe that's per SE (so, 4x128kB maybe), with the per-CU cache now dubbed L0? Guesses, guesses...