Recent content by neeyik

As one of the very few people who bought the actual game (reason: curiosity, impulse purchase, early onset of senility, take your pick) and owner of almost every other LOTR-licence title, this is pretty much the truth about Gollum. Games based entirely on stealth are fine, as are those heavily...

It somewhat depends on how frequently the voltage is dropping below 12 volts and what the power consumption is at that time, but I personally wouldn't be too worried about 11.9 volts -- I'd give the connector a thorough check, just to make sure that all the surfaces are completely flush with...

<p>There are few things <a href="https://www.techspot.com/article/2660-the-dark-web/">as infamous as The Dark Web</a> in the world of technology, but what is it, and why does it have such a bad reputation?</p> <p><a href="https://www.techspot.com/article/2660-the-dark-web/"><img...

Apologies for that -- we changed the chart to better show the differences between the various CPUs, as Hogwarts Legacy was somewhat of a mess at the time of those reviews. Can officially confirm that you haven't lost the plot. It didn't show it with the original chart, though -- AMD never...

I'm not convinced that converting jpegs into a DX BCn format, be it done on hardware or otherwise, would be a welcomed idea, in this era -- neither are lossless algorithms, after all. With the rest being tga. S3TC could be enabled with a bit of console toggling, giving relevant GPUs a nice...

Those figures are normal. What you want to look for is the voltage consistently dropping below 12 volts. If the sensor is constantly reporting, say, 11.8 volts at full load, then there's something amiss.

TSMC’s N6 offers an 18% die density improvement over N7, but nothing extra in terms of performance and power. Navi 23 is 46 million transistors per mm2, Navi 33 is 65 which is a 41% increase in density — that’s far more than just a simple node change. So AMD has clearly done some serious...

Data read/writes from/to DRAM is by far the slowest aspect of any part of rendering -- this is why GPUs have so many cores, so that they can continue to work on new instruction threads while previous ones are stalled waiting for data. Adding any additional latency to the memory system is...

@KofeViR @m3tavision -- take it to private messages, please.

It could well be a combination of both factors, but given that other RDNA 3 GPUs seem to have this issue as well, the problem is one that AMD really needs to address somehow. Take the 7900 XT, as an example -- compared to the 6800 XT, it's 32 to 35% faster, on average, with some titles being...

What you gain with memory footprint by using something like jpegs is then countered by the fact that it's not a block compression. Texture sampling is always done with a 4x4 pixel kernel, which is the same size as the blocks used in BC1-BC7, thus helping reduce bandwidth usage. If the...

The small advantage the RX 7600 has over the RX 6650 XT (5% on average, at 1440p) leads me to suspect that there's seriously something wrong with AMD's compiler - the new card has roughly double the peak FP32 throughput, so even though all the other metrics are the same, it should be performing...

Specs for cards in review (sorry for being late with this):

Average fps @ 1440p, taken from the various tests Steve has done with these cards and then averaged again, against the MSRP launch price. What can one take away from this? Well, perhaps if Nvidia had used the same pricing strategy with the 4090 and 4080, as it did with the 4070Ti/4070/4060...

FH5 is a good poster child for RDNA 2, as seen in our 7900 XT vs 6800 XT article. It's perhaps just one of those games that the RDNA 3 compiler does a poor job of it. As Nvidia, it's been using the game as a flag-waving exercise for DLSS 3 so there are probably some FH5-specific optimizations in...