Have you seen any reviews where they test the efficiency of the SMT (not just in games)?
I've looked, but not seen anything.
Only certain games have shown negative bias so far as dealing with SMT is concerned, nearly every other application, whether lightly threaded or not, has shown gains from AMD's version of SMT better than even Intel's HT which btw is over five gen old.
The games themselves could be better equipped, with patches, to handle SMT & I believe there'd be a dedicated patch for Windows (OS) as well.
You'll find Stilt's review in this very (sub) forum, then there's guru3d just from yesterday ~Have you seen any reviews where they test the efficiency of the SMT (not just in games)?
I've looked, but not seen anything.
You'll find Stilt's review in this very (sub) forum, then there's guru3d just from yesterday ~
Yes & there's a possibility of an OS patch being released sometime next quarter, I suspect SMT gains could be even higher across the board.Cheers. I'd checked out the guru3d reviews, but they only tested SMT off with games (like most of the reviewers).
Overall from Stilt's review it looks like the SMT implementation is very good outside of gaming.
Only LINPACK and Euler 3D were negatively impacted, and in most tests it outperformed HT in percentage gains.
I haven't seen such results outside of Joker's results.Yes, there is no doubt that they went heavy on the negative side and failed to be even-handed.
But, they are called Gamer's Nexus... Ryzen 7 just isn't a hardcore gaming CPU. If you look at the world through that narrow prism (and only ~20% of the HEDT market...) then a CPU like the 6900k or 1800X will not be appealing.
Still, we have numerous reviews showing gaming results actually being in-line with the rest of the chip's performance. They all have a few things in common: they are running updated BIOS, fully updated Windows 10, focusing on averages more than maximums (who really cares if you can get 30% higher maximum when your average is only 5% different?), are configuring Windows power settings properly, and aren't using Asus motherboards (not to dig at Asus, I have full faith that they will work diligently to resolve the outstanding issues and deliver a fantastic product - which is why my C6H order still stands ).
Well there's this ~ http://www.overclockers.com/amd-ryzen-7-1800x-cpu-review/I haven't seen such results outside of Joker's results.
1800X constantly falls short of the 6900K in gaming, barring outside forces like the 200FPS cap in DOOM.
I don't expect the 1800X to keep up with the 7700K in everything, but I do expect so when compared to the 6900K. Computerbase's test suite for example clearly favors the 6900K over the 7700K, and yet the 1800X is sitting below the 7700K there.
The Ryzen 1700 looks like a great value chip.
The only result I've seen that bothers me is the Watchdogs 2 results.
I wouldn't expect Ryzen to be able to make up the IPC and clock speed deficit to the 7700k on minimally multithreaded games/applications.
Watchdogs 2 however is multithreaded to the point where the 6950X shows a significant performance lead if the GPU isn't bottlenecking things.
Yet the Ryzen processors (while still out performing the 7700K) are losing out significantly to even the 6800K, let alone the 6900K.
This makes me wonder if the SMT implementation has some serious flaws.
The idea of having to check individual applications to determine whether they are helped, hindered or unaffected by SMT is very unappealing.
Review shows 1700x ~ 7700k with gtx1080
1080p very high no aa
Well there's this ~ http://www.overclockers.com/amd-ryzen-7-1800x-cpu-review/
The GPU tested here was 980 Ti. Other reviewers were using 1080's or even Titans, which further moves the bottleneck over to the CPU.Well there's this ~ http://www.overclockers.com/amd-ryzen-7-1800x-cpu-review/
I think with bios updates, windows updates and game engine updates in the very worst cases r1800x will be close enough to the 6900k for it not to matter.
In 4k gaming which is likely the type of resolution people with 1000$ cpus will be looking at will not make a difference.
So reviewers with their induced bottleneck that has relevance to very few users, messed up in a huge way.
Right, my point was most people are going to be GPU bound with powerful cpus such as these.It's wrong to insist that 1080p sucks and high res is fine.
1080p is FINE, if you game with a GPU fit for it.
It's also wrong to assume that most that buy Ryzen would buy some high end GPU as most don't.
Most Ryzen buyers will get a 470 or 1060 and game at 1080p or some even push to 1440p.
Seems that Ryzen has some kind of latency problem and testing at low res hugely harms its results in gaming.
On the other hand, at sane res for the GPU, the compute available helps it perform as it should.
So reviewers with their induced bottleneck that has relevance to very few users, messed up in a huge way.
Ofc if there is a fix, it would be nice to make it even better as it would increase perf at any res.
Well the induced bottleneck is used to show how high the CPU performance is. Of course almost nobody will use it in 720p, but at the same time, who will buy it to run synthetic benches all day long?
There are several reviews with better bios that increases performance by more than 10%, close enough that its not going to matter, especially when you can buy 2 1800x for 1 6900k.Even at 1080P high, as long as the GPU isn't the bottleneck the performance gap looks about the same as that computerbase.de review at 720P.
Gamernexus and Techspot/HardwareUnboxed got much the same results.
http://www.gamersnexus.net/hwreview...review-premiere-blender-fps-benchmarks/page-7
http://www.techspot.com/review/1345-amd-ryzen-7-1800x-1700x/page4.html
Steve at Techspot tried Asus, Gigabyte and Asrock boards as well.
Even at 1080P high, as long as the GPU isn't the bottleneck the performance gap looks about the same as that computerbase.de review at 720P.
Gamernexus and Techspot/HardwareUnboxed got much the same results.
http://www.gamersnexus.net/hwreview...review-premiere-blender-fps-benchmarks/page-7
http://www.techspot.com/review/1345-amd-ryzen-7-1800x-1700x/page4.html
Steve at Techspot tried Asus, Gigabyte and Asrock boards as well.
IF I ever receive my Asus ROG Crosshair VI mb from Amazon, I will be using 2 RX480-8g reference gpus in CF with a Ryzen R7-1800X and will run some benchmarks. I'll keep you posted.
Ryzen has consistent lower gaming performance vs 7700k on nonwindows platform as well.
http://www.phoronix.com/scan.php?page=article&item=amd-ryzen-gaming&num=1
Does Linux have SMT scheduling issue like windows too? Switching CPU governor from ondemand to performance in the Linux kernel does boost ryzen performance quite a bit interestingly enough.