51–60 of 67 posts
As someone not well versed in GPU and CXL, would someone mind explaining the significance of this.
Why did something like this take so long to exist? I’ve always wanted swap or tmpfs available on old RAM I have lying around.
I'd have rather a question why we had single (or already dual) core CPUs with dual-channel memory controller and now we have 16-core CPUs but still with only dual-channel RAM.
Earlier quoted context omitted.
I have never had to go deep into NUMA configuration personally but couldn't it be leveraged here?
Yes, if you want your app to be aware of CXL you can configure it as a separate NUMA node.
Do like the card though, was waiting for someone to make an affordable version (or rather: this looks affordable, I hope it will be both that and actually obtainable. CXL was kinda locked away so far…)
> Buy From One of the Regions Below > Egypt :-/ But, because I'm a good sport, I actually chased a couple of those links figuring that I could convert Egyptian Pound into USD but https://www.sigma-computer.com/en/search?q=CXL%20R5X4 > is "No results", and similar for the other ones that I could get to even load
That is pretty hilarious. I wonder what’s the reason behind this. Maybe they wanted plausible deniability in case someone tried to buy it (“oh the phone lines were down, you’ll have to go there to buy one”).
> Buy From One of the Regions Below > Egypt :-/ But, because I'm a good sport, I actually chased a couple of those links figuring that I could convert Egyptian Pound into USD but https://www.sigma-computer.com/en/search?q=CXL%20R5X4 > is "No results", and similar for the other ones that I could get to even load
That is pretty hilarious. I wonder what’s the reason behind this. Maybe they wanted plausible deniability in case someone tried to buy it (“oh the phone lines were down, you’ll have to go there to buy one”).
CXL is going to be really interesting. On the positive side, you can scale out memory quite a lot, fill up PCI slots, even have memory external to your chassis. Memory tiering has a lot of potential. On the negative side, you've got latency costs to swallow up. You don't get distance from CPU for free (there's a reason the memory on your motherboard is as close as practical to the CPU) https://www.nextplatform.com/20…
I would be more worried about memory bandwidth. You can now add so much memory to your servers that it might take minutes to do a full in-memory table scan.
CXL is going to be really interesting. On the positive side, you can scale out memory quite a lot, fill up PCI slots, even have memory external to your chassis. Memory tiering has a lot of potential. On the negative side, you've got latency costs to swallow up. You don't get distance from CPU for free (there's a reason the memory on your motherboard is as close as practical to the CPU) https://www.nextplatform.com/20…
The latency concern is completely overblown because CXL has cache coherence. The moment you do a second request to the same page it will be a cache hit. I would be more worried about memory bandwidth. You can now add so much memory to your servers that it might take minutes to do a full in-memory table scan.