Jump to content

SteelSkin667

Member
  • Posts

    20
  • Joined

  • Last visited

Awards

This user doesn't have any awards

SteelSkin667's Achievements

  1. The memory bus width does have a very significant impact on memory performance, as a wider bus is often the main way to scale memory bandwidth up when GPUs in a given generation only get to use one or two types of memory. Those figures are both gigabits per second, but the 14 Gb/s one is "per pin" (on a hypothetical 1-bit bus), or how many transfers the memory can do per second. If they used the same notation as DDR, it would be 14000 MT/s. On a 256-bit data bus, 256 bits of data will be transferred at a time, which is why you then multiply that per-pin figure by the bus width to get the actual memory bandwidth of the graphics card.
  2. The specs for DDR already take into account the fact that is is double data rate, hence why it is expressed as MT/s rather than MHz. So the bandwidth of dual channel DDR4-3200 would be 3200 * 64 * 2 = 409600 Mb/s or 50 GB/s. However, its actual clock speed is 1600 MHz.
  3. Usually GDDR specs use the Gb/s per pin notation (that is gigabits per second, not gigabytes), which already takes into account the differences between the different types of memory. All you have to do is to multiply by the bus width. For example, the PS5 uses 14 Gbps GDDR6 memory on a 256-bit bus, so the calculation simply goes 14 * 256 = 3584 Gbit/s or 448 GB/s (which is the figure stated in the specs). The relation between the bandwidth per pin and the actual underlying clock speed of the memory does change depending on memory type, as newer types of GDDR will transmit more bits per clock cycle, but usually the only time you'll see it is if you are overclocking your VRAM.
  4. The PCIe situation is a bit puzzling. That M.2 slot in the PS5 is exposing 4x Gen 4 lanes, and then that integrated SSD is also connected through another 4. Presumably once the chip relocated to another board these eight lanes would still be available. Maybe some of the I/O had to be disabled at the same time as the GPU?
  5. They didn't do any math, Linus is just reading benchmark results. The way you would calculate memory bandwidth is by multiplying its data rate (usually expressed in MT/s or in Gbps per pin) by its bus width (64 bits per channel in the case of DDR type memory).
  6. I was also a little bit bothered by the weird ray-tracing / rasterization dichotomy, when the majority of current titles use both techniques; Even that new version of Metro Exodus, which is using RT for the entirety of its lighting, is still rasterizing geometry. Two notable exceptions are Quake II RTX and Minecraft RTX, which are actually path-traced. I believe the only part of these that is rasterized is the UI.
  7. There is no lack of optimization in current games, 'powering through' has always been a thing in the PC space due to the open-ness platform that makes optimizing for specific hardware impossible. Hell, I'd argue that with the advent of Vulkan and DX12, developers are putting more effort in the engineering of their games engines than they ever did. The current setup where RT is an optional extra is just a compromise due to RT-accelerated hardware still being relatively new. In an extra 3-5 when it becomes reasonable for developers to ask for an RTX or RDNA2 video card as a minimum specification, most games will probably either use ray-tracing as an integral part of their rendering setup (ie you can't turn it off) and have their content set-up for an RT renderer, or not make use RT at all. It'll become as much of a technical decision as an artistic one (and there is a very fine line between the two). The GPU price creep is real, though, and it is a major bummer.
  8. Including Metro Exodus or Control to the lineup of games instead of SOTTR might have made it easier for the candidates, although SOTTR is good to include as a shining example of a game where RT barely brings anything of interest to the table - which to be fair that is the case of most of these first-gen games. Andy trying to compare reflections in a game that doesn't have RT reflections was pretty funny as well. The bottom line is that the more of the rendering of the game is being handed off to RT, the more it contributes to the final image. Those early games where rays are being traced just for shadows, or just for reflections are the ones where the trade-off to performance is generally not worth it on present day hardware. Also, nice demonstration of the placebo effect. Poor David.
  9. I find it astonishing that Apple are now using the same M1 chip (well, two variants of it anyway) in five separate machines; they really are making the most of the R&D costs involved in developing these. As for the new iMacs, it's fair to say that they are pretty much overgrown iPads at this point (albeit without touch), but I'm sure they will be very nice to use. I just think that they should have opted for matching colors for the display bezel - the white is kind of ugly imho.
  10. That was mildly entertaining, but now I wonder what the result would be with some more time and actual thought put into it.
  11. Oh, I misread, sorry about that. Yeah, that icon rings a bell! I don't have my 475 anymore (like so many retro computers mine got thrown away by my parents when I left home), but I seem to recall the cache to be on by default - you had to disable it for compatibility.
  12. The 475 came with a 68LC040 and not a regular 68040, so it doesn't have an FPU to enable. That's part of what differentiated it from the higher-end stuff.
  13. My very first computer way back in the day was a Macintosh LC475, so watching this felt quite nostalgic. I remember watching videos on that machine that actually played smoothly (by that I mean "as intended", since they were probably all encoded frame-rates under 15 fps), but resolution was much lower than 320x240; that would have been A LOT of pixels back in the day. It was also an entry-level model (actually the bottom of the barrel at the time), so color depth was super limited and playing games on it was rough, especially after the Power Macintosh came out.
  14. Yes. I've seen some people online buy some P106s for really cheap (they were going for as low as $30 a pop just one year ago) to put them in folding rigs.
  15. I'm as frustrated as you about the current pricing, and I really think no one should give in to the current scalper prices as it only encourages them, but the solution to me is definitely to wait it out. I was planning on an upgrade, but scalpers can FOAD so I'll just wait as long as it'll take. However, when you say that "t nothing new runs well on a 1060" that is blatantly false, it is still perfectly fine and will run everything decently well. You can and should wait it out with the rest of us.
×