It's obvious that the games that show smaller improvement over Zen2 are not that CPU but GPU bound at 1080p, this is reflected in the charts versus 10900K. Zen3 will be a better gaming chip than 10900K, there is no doubt about that, CPU bound games like LOL and CS GO confirm that (with a massive 20% advantage for Zen3).
And when getting pointed to another reference that didn't suit your narrative, your answer was: I don't know about that.
Well I KNOW, the argument was brought up because we were talking about CPU-bound games, then Mr. Tamz clearly misunderstood somebody and decided to present Doom Eternal as CPU-bound, which it is absolutely not. Then that escalated through some people. The point was this all along: hiding behind falsely applied statistical principles, @tamz_msc discards numbers from actually CPU-bound games, then goes ahead to lecture the whole forum 5 pages long about statistics in one of the most distasteful style I've seen in a while - for literally nothing, because his reasoning of deciding whether something is an outlier or not is totally misguided.
Yes, they stated this more than a year ago too (Mr. Papermaster in an interview to AT), and in this presentation as well.
That made me giggle.
We should also not forget that we actually need both companies to do OK performance wise. AMD did sell their top FX CPU for $1000 for multiple years. Nothing could touch them, and there were budget options that trounced Intel as well, that's also true. But still, the $1000 top FX CPUs were out of line in my opinion.
How many negative titles do you want to see in an AMD gaming slide?
1. I ran a simple computation on a given set of data that is publicly available - and that is me being selective about it? If you have doubts regarding the results I obtained, feel free to do the computations yourself.
Why shouldn't we talk about custom memory scaling? AMD took the liberty to demonstrate their CPU as well as the Intel one with out-of-spec memory, so it is only natural that both should be judged on how well they can maximize their potential in this front.
For enterprise, the Zen 3 counterpart to the G series will come, probably early to mid next year after mobile Zen 3. Lot of folks seem to just forget about the Renoir G Pro desktop APUs just because they're OEM only.
Neither you nor Mr. Statistcian moderator from the pro-AMD camp has been able to convince me as to why my application of statistical principles is wrong.
I think the slides are fine and considering it's marketing, fairly balanced. You just asked a question and I presented the counterpart, meaning just because they're not showing Doom Eternal, it's not necessarily because of a weak example. Because that was the earlier question you answered to.
Another ill-willed comment for no reason. When have you seen Mark moderating on any of the arguments he was a part of? This Pro-AMD camp also gets me every time, because when I point out BS from AMD elsewhere in an environment where real AMD fanboys are present, I get trounced as pro NVIDIA etc, not to mention I don't even have any AMD hardware in my computer. As to your statistical principles, I totally get it from you don't worry. See my other comment: at least you're wrong with a conviction.
From your own link going from 3600CL16 to 3733CL14 on Zen 2 leads to a 2% gain, while on Skylake going from 3600CL16 to 4133CL17 leads to a 13% gain. So Skylake gains more from memory overclocking than Zen 2, as for Zen 3 the jury is out there as to how much it can gain from overclocking memory, but we know that the IO die is the same, meaning that no improvements have been made to the memory controller, so chances of Zen 3 suddenly becoming better than Zen 2 in memory scaling don't seem very high.
I was wrong about Doom Eternal, but you still haven't shown me how I am wrong with applying statistical principles. As far as the opinions of the moderator is concerned, it is irrelevant as he basically dismissed everything with an appeal to authority.
No need for him to back it up, AMD officially confirmed the and I/O die and therefore the memory controller to be the exact same, though not in the presentation itself. In press calls that happened just before the stream, in particular to Gamers Nexus and Ian from AT. They've also said, that in this generation 3600-3733 will remain the sweet spot. That's probably why they showed tests using 3600 RAM.
Why don't you try the base supported frequency to the fastest they can run, and check those percentages. You are comparing apples to oranges. In your example about, you have the Intel going to a higher number than AMD.
Yes I did. You are dismissing examples that are, because of their characteristics (being CPU-bound), more relevant to what you are debating (gaming performance of the CPU). You can twist your nonsense another 5 pages long, all that does is just strengthen your opinion as an outlier, thus dismissible.
while 2.3% and 3.4% are small numbers, the total difference in gaming performance is not much different than that. But it does prove my point that AMD gains more than Intel with faster memory. Also, all of this will not matter with Zen 3, as its a whole new architecture.By the way, going from 3200CL14 to 3600CL16 means a 2,3% on Intel system and 3.4% on the 3900X. Not what I call a groundbreaking difference. Also, comparing very higly clocked RAM is not meaningful (as Ryzen 3000 series has the divider in IF clock after hitting 1800MHz). Gamersnexus has also some scaling for very high clocked memory and overclocked IF bus. For Ryzen 5000 differences in theory should be lower as the unified cache pool should limit more the need for memory access.
Ryzen 3000 Memory Benchmark & Best RAM for Ryzen (fClock, uClock, & mClock) | GamersNexus
stub Memory speed on Ryzen has always been a hot subject, with AMD’s 1000 and 2000 series CPUs responding favorably to fast memory while at the same time having difficulty getting past 3200MHz in Gen1. The new Ryzen 3000 chips officially support memory speeds up to 3200MHz and can reliably run...www.gamersnexus.net
How can you come to that conclusion with ONLY that data?
Facebook
Twitter