Recent content by hkultala

  1. H

    SMIC Move To 7nm Node

    "7nm" and "12nm" are only marketing names. The number in the marketing name tells absolutely Nothing about any true metrics of the process. The most important metric of manufacturing process is MMP. TSMCs N7 has MMP of 40nm, GF's smallest processs have MMP of 64nm. But the MMP of SMICs...
  2. H

    AMD RDNA3 Specifications Discussion Thread

    Navi 21 was not successor of Navi 10. Your comparison is totally skewed. Navi 21 was MUCH bigger and MUCH more expensive chip which had over TWICE the area and much more than twice the mfg cost. Navi 22 was the successor of Navi 10.
  3. H

    Speculation and Rumors: AMD RDNA4 ...

    What do you mean by "completely revamped"? If you mean totally new architecture, then you were just wrong. YOU WERE WRONG. Not AMD. But dual-issue of FMA is anyway a significant change. It's not a thing that can be just "tackled" in. It does mean huge changes, even though it does not mean...
  4. H

    Speculation and Rumors: AMD RDNA4 ...

    Architecture disappointing? I think you got things totally mixed up. The architecture got much more improvements than expected, but because the leaked numbers were misinterpreted and bad speculation was based on those misintepreted numbers, people had unrealistic expectations. And then when...
  5. H

    Nvidia GeForce RTX 4080 Reviews

    As long as the important branches of the BVH tree fits into the caches, ray tracing should not need very much bandwidth. But the 3dmark raw ray tracing test result really does look weird. The port royal test, on the other hand, does not look suspiciously low.
  6. H

    AMD RDNA3 Specifications Discussion Thread

    Nothing weird in this. The number of raytracing untis only increased by 1.2x and they got 8-9% more clock rate => 1.3x raw theoretical raytracing performance compared to RDNA2. That they still got in average ~1.6x improvement means that their architectural improvements are giving in average...
  7. H

    AMD RDNA3 Specifications Discussion Thread

    No. It's "up to 1.5x". Marketing speech. I average the performance increase is much smaller. The 1.5x comes in situations where RDNA2 had some bottlenecks and did very badly and was very far from it's theoretical performance. Now RDNA3 gets it's real-world performance closer to the theoretical...
  8. H

    AMD: RDNA 3 Speculation, Rumours and Discussion

    What bugs are you talking about?
  9. H

    AMD: RDNA 3 Speculation, Rumours and Discussion

    Thinking about "raster performance vs RT performance" is the wrong way of thinking. It's really about shader performance vs raw rasterization performance vs raw RT performance. And shader performance is needed for both rasterization AND ray tracing. And shader performance has always been...
  10. H

    Nvidia DLSS 3 antialiasing discussion

    DLSS3 just makes input lag worse, so it's just bad for any competitive gaming. That it gives more FPS is irrelevant when those extra FPS comes at cost of worse input lag.
  11. H

    Apple (PowerVR) TBDR GPU-architecture speculation thread

    "32 MB TLB" is seriously wrong terminology. Each TLB entry takes something like 8 bytes, so 32 MB TLB would have about 4 million entries. With 16 kiB page size, those 4 million entries would mean TLB that can hold the address translation for 64 gigabytes of memory. Which is insanely lot. So...
  12. H

    AMD Radeon RDNA2 Navi (RX 6500, 6600, 6700, 6800, 6900 XT)

    No. When a memory bus is idle, it does not consume (significant amount of) power. And the power spent for refreshing the dram arrays is not greater with gddr6x than gddr6.
  13. H

    AMD: RDNA 3 Speculation, Rumours and Discussion

    And all these power/area and cost/area increases just means MCMs ("chiplets") are making even LESS SENSE. Because that MCM which has lots of die area is VERY expensive to make and very hard to cool. The amount of chip area that we can effectively cool inside one package and which consumers can...
  14. H

    AMD: RDNA 3 Speculation, Rumours and Discussion

    ... so you would have multiple separate clusters of ROPs and long latency between them. Quess what happens when you render two overlapping triangles, one in another die and and another on another die almost at the same time? You either have cache coherency between your L3 caches (which always...
  15. H

    AMD: RDNA 3 Speculation, Rumours and Discussion

    Moores law has never been about performance. It has always been about the number of transistors per chip. And that 3080 has 2.33 times more transistors than your 1080ti, and ot came about 3.5 years later. That's about doubling per 3 years, only slight slowdown in Moores law. And about...
Back
Top