Yeah, its also the way the 5900X and 5950X are identical, usually there is a couple % between them, its a 2080TI.
-
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.
Just watched Linus's performance results of the 5000 series
- Thread starter rn2
- Start date
Yeah, its also the way the 5900X and 5950X are identical, usually there is a couple % between them, its a 2080TI.
linus used manually tuned 3600 cl14
I would be interesting to see what is happening within the Nvidia driver.
Nothing, i just think the GPU is maxed out.
Caporegime
- Joined
- 8 Jan 2004
- Posts
- 32,586
- Location
- Rutland
And the 5900X (will have one decent 6 core CCD and a lesser 6 core CCD). Which is why the 5900X isn't much more than than the 5800X.
Didn't he also have MCE disabled on Intel?.
When I say that what I mean is that there's an inherent limitation to games where you are always going to have your other cores wait around for the main core to get ready for the next batch. The extent obviously will vary on a per-game basis because it very much matters what's actually happening & what other work is being done and kept around in the 'background'. So you'll never really have a game fully utilise your CPU because the design is the game comes first and then we look at how the tools (CPU etc) help us achieve that vision, rather than "let's design a game that uses 100% of the CPU". It also happens with GPUs all the time but it's generally not apparent unless you're savvy about it and can notice it, eg with how Vega was (under-)utilised, or Ampere now at lower resolutions because even GPUs can't simply use 100% of themselves in games there will still be variations in workload type etc.
So while I wouldn't say that an 8-core CPU won't have an advantage at all thanks to consoles, I wouldn't accept the framing of the post I was responding to, which was 'will the 5600x now be dead in the water 'cause of 6 cores'. And that's mainly what I was getting at. In real world scenarios those extra 2 cores won't make a massive difference in games because 6 is already quite a few and there's CPU limits to the way games are coded, and there's also every greater difficulty to properly use more cores (diminish returns).
Thus a 5600x is a great buy regardless of console specs.
Soldato
- Joined
- 6 Jun 2009
- Posts
- 5,460
- Location
- No Mans Land
Last edited:
Agreed
But still not very representative since no one buys a K SKU and then runs it at stock which is why I think hardware unboxed is a more realistic comparison.
Imo all CPU reviews should show the stock and overclocked numbers especially since overclocking is a feature on Intel you pay a premium for.
If they are not going to then they might aswell just use Intel non K SKUs in the benchmarks and use their much lower price points in the comparison.
So tuned up the memory yet left the CPU stock...
Last edited:
Just as well I ordered a 5950X then.
even the 5950x wil
Other benchmarkera tested CPUs with overclocks if you are that way inclined and actually most pc gamers don't overclock their cpu even if it's a k model.
This, it even (marginaly) bests the 3900x in some games.
The only downside, is for some odd reason the reviews show it running a bit hotter than the 5900x.
No doubt bang for buck is the 3600x, and the 3900x/ 3950x are the productivity kings. Ya pays ya money ....
Enabling MCE is just a setting like XMP so I'm sure most would do this and probably more than would go to the trouble of fine tuning the ram.
I dunno man take it up with Linus that's how he does all cpu tests. To be fair it's a level playing field, he didn't overclock the ryzen cpu either - if he turns mce on then he should turn pbo on for the ryzen
Well it isn't crap, just has a couple of dud cores.... One could argue the 12 cores are left over crap as well, but bear in mind there are lower core count CPU's as well, how badly did they die
And maybe also run the ram at 4600/16 like you can on Intel.
I mean he set it all up to heavily favour the ryzen CPU and this was why his results showed that compared to other reviewers.
Sure, I don't generally see reviewers use faster than 3600. Speeds of 4000+ is mainly used by overclockers trying to set records and require high voltages that people don't want to use on a daily machine - it's also very expensive to buy.
No he did not set it up to favour anyone, he used the same ram speed on both systems.
atop looking for excuses, intel lost just like trump lost, just accept it
Stock vs stock is a fair comparison. Overclock vs overclock brings factors such as mobo quality and capabilities as well as CPU silicon quality into question. The kind of performance is then not guaranteed to us mortals who are watching these videos.
I am not aware that HB use Intel systems with MCE turned on. If they did then that’s fairly naughty of them. Although with the way ryzen works - PBO is effectively an AMD overclock.
memory turning is fair enough. 3600C14 you can get standard XMP kit off the shelf. If they were using 4000 with 1.45V for testing and running c17 then that would be extremely dubious as that is fairly specialist ram and not even the best ram can guarantee that sort of the performance as you got to be pretty lucky.
Sorry but I don't follow your logic. AIUI there are simply two 6-core CCDs. Not one high quality and one low.
He's put out another video covering this and apparently it's down to his using 3600 MHz RAM whereas others were using 3200 MHz.
Apparently his results are consistent with other testers using DDR4-3600.
I think he said Derbauer got DDR4-4000 working with 2000 MHz infinity fabric.