When he says it's better, he's talking about 5%. Also they just said with Raptorlake release the 2-4% improvement with prefetcher and cache changes finally got it to Skylake level. Obviously even Skylake has 2x vector performance so they aren't talking about that.
Vector performance only...
No the difference is greater than that. The overall difference is 40-50% between Raptormont and Raptor Cove. Remember, Gracemont is uarch Skylake-level, but it has a weaker uncore. Raptorlake's advancements brings it to Skylake level.
Sunny Cove = 18%, Golden Cove = 19%, makes it roughly 40%...
Intel Newsroom. It has multiple high resolution photos, including the relevant one.
12 cores could work too but 8 core x 18 cluster just fits so well. We don't know the details of Sierra Forest and whether they are using the same client core as Crestmont or even adding enhancements. Unless they...
Actually if you download the images from Intel's site and zoom in, it's pretty clear to me that it's 4x4 config with two tiles missing, meaning there's only 14 clusters.
The bottom rows do not exist. In certain angles it looks that way but that pattern is repeated through the wafer, where the...
Yea and the weird split we have now between product lines should all converge in that year with Panther Lake.
It's ONE compute tile. Where are people getting two compute tiles? Two compute tiles = 288 cores.
Redwood to Lion Cove is probably close to the norm, but desktop is going to skip Redwood Cove at least for the high end, so it'll be able to claim greater than normal gains, just like Core 2 was said to be a near 2x gain while the real predecessor was Core Duo and was only about 20% faster...
Zero. Server chips always take longer to come out so Lion Cove would be too early. With Lion Cove we're talking potential 30% gain over Golden Cove. Redwood Cove itself is maybe 3-5% so vast majority is Lion Cove.
If Granite Rapids was Lion Cove, they could keep core count and frequency the...
I think Koduri being a problem is just a consequence of the real problem that has been plaguing Intel for the past two decades, culture and management problem.
If they were actually on the ball, then they would have been better long time ago.
Ah well. Based on what @Exist50 is saying Koduri's...
Boris, I don't know where you got that. Look at it closer. It's ONE compute die. Granite Rapids has 1, 2, and 3. Whoever tells you otherwise is wrong. The rest are IO dies.
Yea the Enterprise workloads are THAT different. What I believe they did wrong started with Skylake - by unifying E5 and...
Sierra Forest is further different, as the cores are all on on-die along with the memory controller, unlike Rapids which have multiple core tiles.
Yes, there will be crosstalk, but the ability to access caches from other cores are still an advantage over one that can't.
Most workloads won't...
What? Lockdowns increased demand because people were forced to stay at home all day. Again, look at the numbers, Intel gained desktop share.
It decreased after it was rescinded(rightfully), hence the collapse.
We're spending our free time posting here because we care. Other people will see...
Yes, @coercitiv is right. It's the push to ridiculous power levels that are causing it to fail. They are absolutely pushing the limits in all directions.
Remember when 980 Ti's could be overclocked to the point where it was 20-30% higher performance? Now we barely get 5%? And it can't even...
No one wants a Raptorlake Refresh. We want them to move to the next gen. They are clocked at 6GHz already, you aren't getting much more other than further manufacturer approved OC.
Why do you think it's only the OEMs that care? Intel kept losing marketshare until Alderlake, and reversed the...
You gotta be kidding about Raptorlake refresh making any impact. I know you are smart enough to know this.
Let's put it this way. Do you think a Raptorlake Refresh would be competitive later this year?
It might not be end of the world if the CPU after Raptorlake Refresh isn't until late 2025...
@Exist50 What you are saying is Intel is basically going back to the E5/E7 era.
It made sense, and they should have done that a while ago. There was very clear distinction between the two however. The E7 had massive amount of memory support, but traded it for latency. It had more cores and...
Way, way back Skymont was mentioned for Core lines. Think Nehalem era, maybe Sandy Bridge.
I am asking because 80 cores are not going to be competitive. They need the 120 core setup. A mere 40% gain over the shoddy EMR chip, really? Sure it'll be for -AP, but that's a different class...
They are all experimenting and learning as they go. When you are on the cutting edge, you will make mistakes, but it's inevitable.
And what works for one company won't work for another. Gotta think dynamically and adjust based on your situation.
China has such influence in the American market that a certain American gaming company banned someone from competing just by saying something Pro-Taiwan.
Granted, that particular company had been retards for a while, but it shows you a glimpse of things, to come.
It depends on the definition of what's needed. CPU and GPU is the most demanding part. If the CPU portion takes only 5W you could probably do all at 10W. And between CPU and GPU it heavily depends on what you are running.
I can tell you the PCH on my Kaby-Y doesn't use more than 1W at max and...
It's engineer, fab guy, or finance. Those are the big choices. If you look at Intel, Fab guy = Barrett, Kraznich, Finance = Otellini, Swan
It seems Fab experience isn't necessarily good for the company because there's a difference between relentless execution and future vision. They excel at...
I didn't notice that. Nice catch. "Sierra Glen"
Well ok, but the original Nehalem was based on Netburst, and they reused the name for the Core-based chip. Also, while a different team altogether, they called successor to Goldmont, Goldmont Plus. Yet the change was substantial and brought 30%...
2021 with the new node naming and announcement of IDM 2.0 and 20A process that Granite Rapids was put as Intel 4.
2022 was when Granite Rapids was moved to Intel 3. Are you sure about what you are saying, that they did not change the core? Since that's a 2021 slide?
I believe you without much question in most things but both Lunar Lake and Royal core claims I take it with healthy dose of skepticism.
Actually I am hopeful for LNL, and my attitude towards Royal is "show me".
Why? First because I've been burnt enough by those claims in the past, and...
Remember how people say engineers should be selected for CEOs?
Well, unlike AMD/Intel Jensen has been CEO from the very beginning of Nvidia and he's an engineer at heart.
Lunar Lake uses the same core uarch as Arrowlake. I bet you it's like Core M saying it's "architected for low power" but it came out to be lot less significant than they said it would.
They made you get pretty excited for not much...
Those slides are old enough that it's pre-change. It was based on Intel 4 and then it got changed to Intel 3+RWC with 10% extra gain. They can't change back to plain-vanilla RWC unless they moved it back to Intel 4.
The very interesting thing is, 80C for SP and 120C for AP is what MLID said...
The changes are apparently so little even the CPUID doesn't change.
If anything you should ask MLID why he believes 10-15% perf/clock gain is correct. Even in that leaked document it talked about a potential overall performance gain, which will not apply for 5GHz+ desktop level chips and high...
That doesn't mean you can ignore CSGO. And LOT of unmentioned DX9 games got a huge uplift.
CS2 is a good thing for Intel because they do fantastic in recent titles. They are firmly at 3060 Ti/RX 6700 level.
Perhaps it's the same thought that managers of the Intel graphics group had when...
Actually the reveal said Otellini was the one that refused the iPhone because he underestimated the volume and per chip price was too low based on those expectations.
StrongARM also had serious development issues within Intel. The problem with Intel is Intel, not the ISA they are using. They...
Can't compare with the Chinese GPU though. Intel has decades of iGPU base. Hence why their compatibility is so much better since they are not a startup starting from scratch.
A GPU startup is unrealistic for that reason, because lots of games need hand-tuning.
This is why Arrowlake is replacing Meteorlake at the high end. Even with 8+16 config(which isn't) essentially releasing Raptorlake again but with less power is going to do bad on desktop.
"Nearly 0 changes" sounds like the gain akin to Raptorlake. 1-3%.
@Hulk So regarding the timeline of Intel...
@Geddagod is saying that Redwood Cove core in Meteorlake is only 30% larger than Zen 4. That's pretty much ISO-process.
And the Crestmont core in Meteorlake is even smaller relative to the bigger neighbour Redwood Cove. Meaning it's 1/3rd the size of Zen 4. It's also likely we get greater...
The IFS guy also has an EE degree by the way. Kraznich had a FAB background, but he was the one that caused Intel to lose the leadership.
Based on the last video I cannot say even Battlemage will reach 95% of the theoretical target. And it'll only happen due to software, not hardware work...
Also remember in 2005 all we worried was about the PC(desktop) side while a tremor was brewing and in early 2007 we got the iPhone. 2006 seemed so good for Intel with Core. The first iPhone released in January 2007, mere 6 months after Conroe/Merom. They released the Atom in 2008. To be honest...
@Hulk
Remember, I am still basing my hopes based on Gelsinger's previous history. Yes, the refute is that one man can't do much, but surely one man can do a lot to ruin it, especially someone like a CEO that has so much power.
He hired TWELVE THOUSAND people, mostly in engineering positions...
Open Source is no magic. Linux, and Ethereum both suffers from the same issue. I always argued that Vitalik Buterin embodied the "decentralized" mindset so much so that it felt like an excuse to be not as responsible, and he take charge until the whole thing develops. If it was a company, then...
TAP said DX11 has to be a labor of love forever. And indeed improvements are coming but due to them laxing on iGPU drivers for the past 20 years, Battlemage is going to inherit the weakness.
You could say it's a "flaw in the hardware" that it's a way different architecture from the conventional...
Unless Battlemage is indeed much more competitive, then in spirit, MLID's former(not the recent one) claims of ARC is more or less cancelled is correct is it not?
He did get few things right such as XeSS name and Redwood Cove core. Someone said he has really good sources, but he misinterprets...
Like I outlined before
Intel 4 - late this year with Meteorlake ULP
Intel 3 - Early-mid 2024 Server with GNR and SRF
Intel 20A - sometime very early 2025 with ARL-P
I don't believe it'll take 4 years because 10nm was a genuine screwup, not just reaching for pie-in-the-sky claims that were...
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.