goldstone77 :
Ryzen 5 review vs. Core i5: Ryzen 5 1600X wins for best mainstream power CPU
We test the best for $250, pitting the new Ryzen 5 1600X against Intel's vaunted Core i5-7600K.
"The problem is, people don’t want complicated answers. They want simple answers and they want you to pick for them. In that case, Ryzen 5 is the way to go. It burns Core i5 to the ground in multi-threaded applications performance and doesn’t give up much in single-threaded performance."
"It’s pretty hard to pass up the incredible performance the Ryzen 5 1600X offers, especially as we move into a world where more cores and more threads are expected to matter. For that new world, the Ryzen 5 1600X is easily the winner and just a hell of a deal for the overall performance you get."
http://www.pcworld.com/article/3186811/computers/ryzen-5-review-vs-core-i5-ryzen-5-1600x-wins-for-best-mainstream-power-cpu.html?page=4
I agree the R5 is the chip for code that scale up well to lot of threads, like encoding and rendering. But PCWorld is recommending the R5 using the same flawed arguments that have been used during last decade. The R5 will be a completely outdated chip before average code scales up to 12T.
I recall when people used this kind of arguments with Bulldozer: Get eight cores because soon a SB i5 will be outdated with future software "expected to" use more threads "now" that we have cheap eight cores and developers can start to use "moar cores". It didn't work.
Latter the argument was changed; soon FX-8350 will be better for gaming than i5 because the new consoles use eight Jaguar cores and developers can start to use moar cores. I remember the Eurogamer article about this, with interviews to game developers predicting how the 8350 would probably be a better choice for future gaming. It didn't work.
Then the argument changed again with Mantle, and then repeated with DX12/Vulcan. Soon eight cores (8T) will be the preferred options for gamers because of the new threading model, they said us. I remember all the expectation and promises, then Mantle was killed, and the first tests of DX12 games didn't match the hype generated.
Much more recently, AdoredTV published videos about how the FX-8350 finally did beat the SB i5 on games
AdoredTV used that to hype RyZen launch, It will be a "beast on games" the guy promised us, but he used wrong data and computerbase review of RyZen did show a different picture with R7s losing to modern i5s and even i3s on games, whereas the Sandy Bridge i5 continues scoring above the 8350 today
https://www.computerbase.de/2017-03/amd-ryzen-1800x-1700x-1700-test/3/#diagramm-gesamtrating-spiele-full-hd
Click on "+29 Einträge" button to see the full chart with FX-8350 and i5-2500k included.
Yes, the gap has reduced from about 17% ahead in 2012 to about 10% in 2017, because games are a bit better threaded now, but that is all.
History is repeating again with PCWorld promising now that R5 will be better than i5 in future because software is "expected to" use more threads. Expected by whom? By the same people that did make the same unfounded promises in the past? My advice is to make a purchase on base to
current performance. Yes, it is good to maintain an eye on future performance for the purchase not getting outdated before one can upgrade to something new, but making purchases in base to unfounded promises about future performance is not something that I would recommend.
UPDATE:
I have checked the full PCWorld review and it is even worse than I believed! The conclusions are based in unfounded claims that the reviewer does about the microarchitecture and about games.
This is one of the relevant flawed parts:
Although there are still some outstanding questions, it’s clear to me that there isn’t some flaw with Ryzen that makes it slow (which everyone feared). The most logical conclusion is to blame the games themselves.
I say this because If Ashes of the Singularity developer Oxide can bump performance by 20 percent or more after a couple weeks’ worth of tweaking, and in fact says it’s not fair to even compare Intel with AMD with the previous code, it stands to reason other games could do the same. Optimization may not erase the difference completely, but it should make any remaining difference insignificant.
Ryzen may still have problems with older games if only because game developers are unlikely to update code for a 2014 title. However, I’d bet few of you are having problems running a three-year-old game with your rig today. A modern GPU and modern CPU can run any older title without issues. The more important question is whether developers will support Ryzen going forward for games that come out in 2020—not 2014.
Conclusion
After testing Ryzen 5, and especially after seeing how its performance changed with optimized games, Ryzen gaming performance is clearly not as big of a deal as it seemed when Ryzen 7 first launched. When it comes to deciding the matter at hand—which is the best $250 CPU—the complicated answer is: Match the workloads above with what you do and choose based on your needs, not what someone tells you is right.
The problem is, people don’t want complicated answers. They want simple answers and they want you to pick for them. In that case, Ryzen 5 is the way to go. It burns Core i5 to the ground in multi-threaded applications performance and doesn’t give up much in single-threaded performance.
On the thorny gaming question, Core i5 still has an advantage for now. We expect newer games will support Ryzen, making the performance difference mostly moot down the road.
LOL. No! The problem is not on games! We cannot blame the games!
RyZen is a microarchitecture optimized for a subset of server workloads. RyZen has been optimized for throughput instead latency. Before launch, CanardPC and me predicted RyZen was going to shine in GPU-workloads but suffer in games precisely because of this reason
https://twitter.com/juanrga/status/836503836260990976?p=p
And reviews just confirmed what we said. Precisely this latency problem is the reason why AMD just released a new AGESA/BIOS that promised to reduce latencies by about 6ns
https://community.amd.com/community/gaming/blog/2017/03/30/amd-ryzen-community-update-2
This latency issue is also the reason why RyZen performance improves more than Intel when overclocking the RAM. Overclocking the RAM reduces the access latency to main memory.
The problem is not on the games. The PCWorld reviewer is wrong. He cites AoTS as a game that improved a lot of after developers patched it for RyZen. What he doesn't tell is that AoTS was performing spectaculary wrong on RyZen before the patch. The game was broken. If Broadwel is about 20% ahead RyZen clock-for-clock on general gaming, the gap was a huge 60% on AoTs
The path has improved 20% the performance of RyZen because the game was broken before the patch. Pretending that all games will get this kind of benefit from patching is wrong and recommending the R5 in the hope that it will game significantly better in future is wrong.
"
Facebook
Twitter
Instagram