By GamerNexus who literally waited like 2 seconds after the presentation to be over to post his video. They are using the same IO, unmodified so it's quite impressive that the IPC gains are without even touching the 12nm IO, the IF speed and connections are the same(not surprising since the IO was built to Server duty standard just a cut down version of the EPYC IO)Where did hear/see/read that?
I was just about to ask about the IF
I think it is the 5600X and 5800X that caused the pushback. I expect 5600/5700X to fix that.
I don't know about the Computerbase test, but here is a summary of what RAM overclocking on Intel can achieve(this was brought to my attention by Buildzoid).It is true that Intel can reach higher memory speeds but Computerbase had benchmarks that showed that all-in-all Intel benefited less from ultra-tuned >4000Mhz DDR4 with custom timings than AMD did from corresponding 3600-3800 Mhz + Custom Timings.
1. The chances of this happening is remote since we have confirmation of the IO die being the same as Matisse.1. You have no idea whether AMDs FCLK or memory scaling has improved (perhaps they can also run at say FCLK 2000-2133 now)
2. As @Kuiva maa stated, BF V really downclocks Zen 2 (to around 4.1 Ghz). If the PBO implementation is better, it might very well help Ryzen in such heavily threaded benchmarks as well. Meaning such all-core OC might help Ryzen as well
Intel was much more than 2% ahead in gaming before Zen 3.When Intel was in the lead in gaming, even 2% was "a big lead", not you are saying 5% is not a big lead ????
Gaming, the last remaining stronghold of a decaying Intel uArch(milking it since Pentium M) it's gone now and some people are grasping at straws..As you can see the gains in 1%lows are absolutely stellar, much more than what AMD can achieve, around 20-30% in most cases barring the presumable outliers.
It ain't gone until independent benchmarks can prove it's gone. Just like RTX 3080 being twice as fast as the RTX 2080.†Gaming, the last remaining stronghold of a decaying Intel uArch(milking it since Pentium M) it's gone now and some people are grasping at straws..
I will let someone else post benchmarks if they want to prove you wrong. Right now you are in denial, and can't admit anything, and I will not spend any more time trying to convince you. Once your mind is made up, its just like beating my head against the wall, and I won;t do that.Intel was much more than 2% ahead in gaming before Zen 3.
So after all this you really want to draw a parallel between the public integrity of Dr. Huang and Dr. Su? I'll have fun watching you try.It ain't gone until independent benchmarks can prove it's gone. Just like RTX 3080 being twice as fast as the RTX 2080.†
†Which was proven to be a cherry picked result by reviewers.
That's confirmed, though AMD omitted it from the video:and on the motherboards, post a link to prove that.
I think it is the 5600X and 5800X that caused the pushback. I expect 5600/5700X to fix that.
I think you got that backward. Several board makers refreshed 400 series boards with bigger BIOS storage and promised Vermeer support for them, and AMD then initially wanted to support Vermeer only on 500 series. AMD then backpaddled. May well be the case 400 series support is late since work on it started later.Ooh that's clever. AMD promised they would support Zen 3 on 400 series, they just didn't say when.
IMO I think that makes the Warhol rumor more likely. Board makers I'm sure wanted Zen 3 to be 500 series exclusive to ensure that they would be able to sell enough during the period between Vermeer's launch and Warhol.
Yeah that 10% is like altogether in the last 5 generations maybe? So, not really the same thing here at all. Zen 3 comes out literally a year and 4 months after Zen 2.Apparently some have forgotten how silly of a model the 3800X was, price/perf wise. But that makes someone an intel fanboy now apparently*. As of Nov. 5th, the 8 core upgrade increase will be $50 (3800) or $120 (3700). For the Ryzen 5 chips its $50 (if we had a 5600 non x) and $100 price increases.
Perhaps in 3-6 months we'll see the 65w variants show up and drop in price to the norms of the past 3 years.
*Oh the memories of the price increase justifications of the skylake thru coffee lake era because like its 10% IPC bro! Moar cores bro!
But AMD should give every feature all the time to older boards too. I think it's fair to expect a company that big to give DDR5 support as well for 400 series motherboards once it comes to market. The least they could do.I think you got that backward. Several board makers refreshed 400 series boards with bigger BIOS storage and promised Vermeer support for them, and AMD then initially wanted to support Vermeer only on 500 series. AMD then backpaddled. May well be the case 400 series support is late since work on it started later.
Personally I honestly don't mind that 2 months gap. For the high end SKUs that launch on Nov. 5th 500 series boards which will also support per core OC will be much better fits. For lower end 5000 series SKUs we'll have to wait longer anyway.
If they support X470 boards it's gonna be hard not to upgrade my 2700X even though I don't really have any complaints about the performance.
Do you disagree that your take is the outlier in this topic?Do you disagree that there are outliers in AMD's gaming benchmark comparison?
No 5GHz = FAIL /s
The announcement looked to be impressive! Looking forward to the official reviews!
Lisa didn't even say the I word one time....What gives?
Date code on the 5900X was 2035 so that's late August.
Price wise I don't really see any issues. I'll be shopping at Microcenter if I decide to upgrade. Would have been nice to see a 5700X offering in the line up.
7nm should now be a lot cheaper than it was when the 3000 series launched, so that combined with the price increases, should see AMD much better off financially in the year ahead.1) Very interesting that they are still on the optimized 7nm process they used for the 3xxxXT chips. Seems they're saving a much bigger change for Zen4.
Only naive people forgetting that Top of The Line CPUs are not budget friendly and Intel Fans are complaining about a price increase. Amd IPC increase from Gen1 to Gen3 Zen net you nearly 40% performance increase while keeping the price of top 8 core CPU less than $500... Mean while Intel was charging an Arm and a Leg for 6 core CPUs from Sandy Bridge to Skylake and their total net IPC gain was like 20% IPC. LOL Yeah Keep calling AMD a Price Tyrant and just another Intel and Nvidia it will show us all how ungrateful you are
Forget grateful... If you don't like the price, don't buy it.. Your loss.Why would I feel grateful to a corporation? Their 6C/12T just went from $200 to $300, why would I not think that sucks and doesn't sound like a good deal?
I'm not in denial - I'm spouting facts. It is fact that AMD used outlier results in their presentation. It is also fact that Intel gains much more than AMD with a memory overclock. And finally it is fact that first party benchmarks are often painting a picture better than it actually is. All these combined leads me to believe that AMD isn't going to win decisively in gaming.I will let someone else post benchmarks if they want to prove you wrong. Right now you are in denial, and can't admit anything, and I will not spend any more time trying to convince you.
Go back to page 9 of this thread - I have proven that there are outliers in AMD's data. If you choose not to believe it, then it's a problem with you brushing facts under the carpet, not mine.Do you disagree that your take is the outlier in this topic?
Meaning it's completely irrelevant, a glitch. According to standards.
Forget grateful... If you don't like the price, don't buy it.. Your loss.