3 ms·

Games for a long time struggled to with scaling up to more then one core. Even today the games scale to like 2 or 4 cores. Just at the PS3 cell, it took a long

by mtanski 11y ago

Games for a long time struggled to with scaling up to more then one core. Even today the games scale to like 2 or 4 cores. Just at the PS3 cell, it took a long time for games to really utilize the PS3's full potential. That's why I don't think that dropping 48 cores on game devs would have been a good business decision.

nickpsecurity 11y ago
You can just drop 48 cores on them or you can drop 48 cores plus tools and patterns to use them well. There's been techniques for linear, reliable scaling in academic literature and even in industry (eg NonStop, MPP's) for years. There were also toolkits that helped automate that process using c-like languages. That said, just because they can't use it right off the bat doesn't mean they won't figure it out over time. We really didn't get to see them try. The Xbox used a non-power-of-2, lame core count while Cell didn't use traditional cores at all. Apple to oranges a bit although 360 is a realistic comparison if the multicore tooling sucks. So, what's Microsoft and Sony's recent approach? Drop a bunch of cores on them. :)
- pandaman 11y ago
  GPU in PS4 has 18 cores, in Xbox one - 12. They work quite well. Not exactly powers of two, but neither is 48, to be honest. The problem with Cell was not that it had multiple cores. It was that it had to do a lot of graphics tasks to take the load off the weak GPU. And doing graphics on a CPU is quite hard. Doing so in concert with another GPU is doubly so. On the other hand, doing CPU work on GPU is much easier.
  - nickpsecurity 11y ago
    Good point. Powers of 2 a bit of a brain fart of mine. That adding more cores worked well was my point in recommending an Octeon-like design. Didn't know the PS3's graphics system sucked that bad. Trying to do a real GPU + a fake one would definitely be hard, esp if the CPU was hard to begin with.
  - mtanski 11y ago
    GPU and CPU cores are very different in how they process data / architecture. You can think of GPU cores as you schedule a series of "vector" instruction over a large chunk of data. They are great at internally paralleling this process because the flow is simple. You assume the there will be little synchronization and little branching. That's why you need such fast memory on GPUs to keep streaming data at instruction speed. CPU cores are a lot more general prepuce where you're generally building the business logic and the work load is a lot more branch oriented. So if you have a problem that's trivially parallizable without (much) branching logic then throwing more GPU cores is easy. Problems like this numerical in nature like large matrix calculations (math or 3d meshes), streaming data (hashing). Hence why it's easier to keep adding GPU cores... it means games can process more geometry data in a fame, but adding more general prepuce CPU cores is harder because the complexity of business logic goes up as you split up the work and have to worry about ordering/synchronization.
    - pandaman 11y ago
      There is not much difference in how the GCN and a generic CPU process data, I can not speak for every GPU out there though. The only significant difference is that each GCN core can run up to 40 threads while the CPU at this power level run at most two (Jaguar in the 8th gen consoles is one thread per core). Having so many threads simplifies a great deal of parallel programming because instead of splitting each task into several chunks, where each chunk still has multiple items that need to be looped over you can just spin a thread for each item and get rid of the loops. Thread switching is zero overhead and creation/destruction is a few cycles. These are not Windows/pthread software threads. Of course, you don't have to run the same code in every thread. If you can figure hundreds of different tasks to do simultaneously in a game - you can put each one of them in a thread too. This is also why a GPUs, generally, use a very slow memory, contrary to your belief, With so many threads latency does not matter that much since a thread that gets stalled on a memory access is preempted for one that already has data available.
      - mtanski 11y ago
        Current gen GPUs use GDDR5 compared to DDR4 in only handful of new Intel chips that started shipping in the last few weeks. The GDDR4 chips runs at 750Mhz and DDR4-2133 as supported by the fasted shipping Intel CPU runs at 266Mhz. That is an effective transfer rate of 48 GB/s vs 17Gb/s for the DDR4. The current GPUs effectively have the fastest off core memory of current devices. They need those transfer rates to keep all the stream processors running.

by mtanski 11y ago

nickpsecurity 11y ago

You can just drop 48 cores on them or you can drop 48 cores plus tools and patterns to use them well. There's been techniques for linear, reliable scaling in academic literature and even in industry (eg NonStop, MPP's) for years. There were also toolkits that helped automate that process using c-like languages. That said, just because they can't use it right off the bat doesn't mean they won't figure it out over time. We really didn't get to see them try. The Xbox used a non-power-of-2, lame core count while Cell didn't use traditional cores at all. Apple to oranges a bit although 360 is a realistic comparison if the multicore tooling sucks. So, what's Microsoft and Sony's recent approach? Drop a bunch of cores on them. :)

pandaman 11y ago
GPU in PS4 has 18 cores, in Xbox one - 12. They work quite well. Not exactly powers of two, but neither is 48, to be honest. The problem with Cell was not that it had multiple cores. It was that it had to do a lot of graphics tasks to take the load off the weak GPU. And doing graphics on a CPU is quite hard. Doing so in concert with another GPU is doubly so. On the other hand, doing CPU work on GPU is much easier.
- nickpsecurity 11y ago
  Good point. Powers of 2 a bit of a brain fart of mine. That adding more cores worked well was my point in recommending an Octeon-like design. Didn't know the PS3's graphics system sucked that bad. Trying to do a real GPU + a fake one would definitely be hard, esp if the CPU was hard to begin with.
- mtanski 11y ago
  GPU and CPU cores are very different in how they process data / architecture. You can think of GPU cores as you schedule a series of "vector" instruction over a large chunk of data. They are great at internally paralleling this process because the flow is simple. You assume the there will be little synchronization and little branching. That's why you need such fast memory on GPUs to keep streaming data at instruction speed. CPU cores are a lot more general prepuce where you're generally building the business logic and the work load is a lot more branch oriented. So if you have a problem that's trivially parallizable without (much) branching logic then throwing more GPU cores is easy. Problems like this numerical in nature like large matrix calculations (math or 3d meshes), streaming data (hashing). Hence why it's easier to keep adding GPU cores... it means games can process more geometry data in a fame, but adding more general prepuce CPU cores is harder because the complexity of business logic goes up as you split up the work and have to worry about ordering/synchronization.
  - pandaman 11y ago
    There is not much difference in how the GCN and a generic CPU process data, I can not speak for every GPU out there though. The only significant difference is that each GCN core can run up to 40 threads while the CPU at this power level run at most two (Jaguar in the 8th gen consoles is one thread per core). Having so many threads simplifies a great deal of parallel programming because instead of splitting each task into several chunks, where each chunk still has multiple items that need to be looped over you can just spin a thread for each item and get rid of the loops. Thread switching is zero overhead and creation/destruction is a few cycles. These are not Windows/pthread software threads. Of course, you don't have to run the same code in every thread. If you can figure hundreds of different tasks to do simultaneously in a game - you can put each one of them in a thread too. This is also why a GPUs, generally, use a very slow memory, contrary to your belief, With so many threads latency does not matter that much since a thread that gets stalled on a memory access is preempted for one that already has data available.
    - mtanski 11y ago
      Current gen GPUs use GDDR5 compared to DDR4 in only handful of new Intel chips that started shipping in the last few weeks. The GDDR4 chips runs at 750Mhz and DDR4-2133 as supported by the fasted shipping Intel CPU runs at 266Mhz. That is an effective transfer rate of 48 GB/s vs 17Gb/s for the DDR4. The current GPUs effectively have the fastest off core memory of current devices. They need those transfer rates to keep all the stream processors running.