-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
Intel’s surprise Ryzen killer
- Thread starter Wulfster
- Start date
Associate
- Joined
- 18 Oct 2002
- Posts
- 1,335
- Location
- Eltham
I wonder what 8.5% of them are thinking.
They just woke from a coma.
What are 33% thinking TBH. 24% of people are buying 2000 series cards...
quite a change huh? soon we will be recommending all amd pcs
I think nvidia will still be on top.. but poor intel..
All AMD PC's, for the first time i think? is actually a thing that has real merit.
An RNDA2 GPU paired with a Ryzen 5000 series CPU, that combination actually gives you a boost, they actually pair up to give you something extra, a benefit to a full AMD system, added performance and performance beyond that of any other combination.
For as long as AMD have been making GPU's and CPU's they had the potential for this, its great now that they can realize it because they are right at the very top with both, for the first time.
remember that this is only if the game supports it... technology originates from the next gen consoles though so it should become fairly common soon
Can you explain that a bit more for me please dude? Just pulled the trigger on an X570 and some sweet RAM, so looking for what's next
Can you explain that a bit more for me please dude? Just pulled the trigger on an X570 and some sweet RAM, so looking for what's next
Infinity Cache or Smart Cache... can't remember what its called now but basically if you pair a Ryzen 5000 series CPU on a 500 series motherboard with an RX 6000 series GPU they can access eachothers Cache systems to give you a boost in performance of 'up to' 13%, in games that support it?
SAM - Smart Access Memory is the feature you are thinking, Infinity Cache is a "Zen 3 like" cache system that reduces bandwidth requirements for the 6000 series.
Mind you via SAM the CPU could well have full access to the cache over fabric. What we need is more techie details.SAM - Smart Access Memory is the feature you are thinking, Infinity Cache is a "Zen 3 like" cache system that reduces bandwidth requirements for the 6000 series.
There we go, thank you "Smart Access Memory"
There we go, thank you "Smart Access Memory"
It's what is also called... ramming your fabric tech down your competitors throat until they choke on it.
It's what's also called ramming your fabric tech down your competitors throat until they choke on it.
It is indeed, Given that at this point at least AMD are the only ones making high end CPU's and GPU and actually innovating instead of regurgitating decades old technology long after it should have been dead..... yes! Absolutely ram it down your competitors throats and make them choke on it.
You remember the intel internal memos? The ones that were like.... "They have a fabric advantage, etc etc?"... This is exactly the sort of tech that comes out of that fabric advantage. I think it's great and compelling while also not being a tie in feature. I like that!
SAM - Smart Access Memory is the feature you are thinking, Infinity Cache is a "Zen 3 like" cache system that reduces bandwidth requirements for the 6000 series.
It certainly opens up some interesting possibilities. One coherent fabric to pull everything together.
My favorite part about all this is that this new stuff opens up new conversation, new possibilities, you're not going down that same road every single time. Now you can explore other aspects and see what is possible by linking outside aspects from the CPU to improve performance. Right now AMD are playing 4D chess 10 moves ahead.
I don't know how Intel got so big... well i do but you look at the history between these two in how they operate, Intel created X86 and with it the core of what became IBM's "Personal Computer" good stuff, well done Intel.
But since then its AMD who have spent their time identifying problems to be fixed and ways to improve what a CPU does and how it works while Intel have been plodding along content with how things are or trying to replicate AMD's innovations only when they have to and badly.
AMD deserves to be on top because they actually work for it.
Can't wait for AMD to revisit their Uniform Memory Access technology as they expand the possibility with Infinity Fabric, this is something they had to abandoned because Intel was setting the direction and didn't like where this was heading, Now AMD have more control again.
Lets get unstuck from the 1990's.
This was with crappy iGPU acceleration, think what this would look like with a high powered decree GPU, its coming, oh its coming...
Lets get unstuck from the 1990's.
This was with crappy iGPU acceleration, think what this would look like with a high powered decree GPU, its coming, oh its coming...
Thanks, very interesting (and geek cool). I did a quick bit of digging and I think the underlying tech has been available (in Windows via WDDM) since v2: https://docs.microsoft.com/en-us/windows-hardware/drivers/display/resizable-bar-support
So it looks like it's a memory mapping technology where the CPU address space (virtual addresses) are mapped to the frame buffer address? I suspect that Nvidia will be able to do this as well without any hardware changes, good to see AMD pushing the innovations though.
I guess that since the ability to do this already exists in the OS, there's probably been lots of "GenuineIntel" compiler shenanigans going on in conjunction with lack of BIOS support for it or we'd have seen it in earlier offerings.
I'm not technically minded enough to understand enough about it but to me it does look like this is the system at the OS level to allow it to work.
However the actual technology in operation is physical hardware on the chips and its AMD's IP, the whole subsystem is called "Infinity Fabric" AMD's CPU's have had it for a couple of generations, they use it to tie physically separate chaplets together into what Windows see's as a single monolithic CPU, AMD's GPU now have the same technology and that's what allows the GPU and CPU's memory architecture being physically separate hardware to behave as one in the way they access each others memory.
Intel nor Nvidia have this technology, if Nvidia or Intel came up with their own versions of it they would still need to be compatible Intel to Nvidia, Intel to AMD, AMD to Nvidia and they would need eachothers permission to access eachothers IP.