Discussion Intel current and future Lakes & Rapids thread

[TheGiant]

Looking at those reviews, the prominent sites did the worst work

You have to watch youtube videos to get more info

I was interested in the new mem controller- that is partially answered with synthetic measurements- i wonder if this new mem controller is a production test of alder lake/later chips with ddr5, where we expect effective speeds up to 8GHz and and IMC probably can't run at that speed

It would be nice to see some exotic mem speed like Ddr4-5000+ at 1/2 mode, not only in gaming but with adobe software or calculations

I can't find any ring speed tuning

My experience with 9900K (2 weeks when a friend was on vacation) is that mem tuning and uncore clock has order of magnitude more impact than clocks, everyone is hunting 5,1 to 5,3 GHz but the gaming min fps limit is not there, it is latency

overall the gaming testing looks on the portals like average gameplay instead of CPU tests- at least Toms did some more demanding test which you see by lower mins within the same settings (like 1080p high..)

gaming CPU test should be composed by 90% of CPU intensive scenes, not running on the grass field or rotating camera inside a city

so mem latency can be 40ns which is like 10% behind the best skylake systems, cache is bigger, clock is the same, IPC is higher- even with those flawed gaming testing scenes, which test more thoughput than randomness I expected more

so about gaming it is a big grain of salt

app/productivity - one should test a machine with its intented purpose- an 8C cpu clocked way above its efficiency point for productivity is like testing a bmw m5 with full loaded trunk and 4 persons pulling 1,5t going up the hill
I mean wtf? that is a side info and for enthusiast a test should be done with decreased clocks and proper voltages
for example I am running my 3900X limited to 65W when handbraking, total performance loss is totally unimportant

power wise- I think the icelake power wasn't only the problem with 10nm, this sunny cove design looks leaky like a broken water dam
I didn't see idle power and lower load power- it is important, the sites don't test workload test packages per day with different distributions of idle, light and full load

iGPU- again wtf? this gpu isn't there to win 3dmarks, but support this age monitors, run everything on every system (win, linux..) and run even the most recent encoded movies with low CPU usage and enable fast decoding even with lower quality when it is not that important
I cant find anything about this

very shallow work from the reviewers, in short run cinebench and be done with it

 

[lobz]

That review shows a much better result for Rocket Lake vs Comet Lake than Ian's review. And somehow they managed to score a state of the art gpu.

Well it's Tom's after all. If there was a chance to show Intel in a better light, that was at Tom's.

That being said, it's kind of hilarious to see all the star reviewers being so mad at Intel, when all they did was release a couple of products at the top-end with practically no progress for a bit more money, and kind of really good products at the mid-range for an attractive price.

 

[TheGiant]

looking at alder lake info
gaming wise I am not so convinced
alder lake looks to have cache structure of tiger lake/skylake sp
we dont know if ring or mesh or what between the small and large cores
ddr5 4800 will have catastrophic latency and will be expensive

gaming wise alder lake doesnt look like a win

but from a business/corporate, this will be a game changer

 

[coercitiv]

That being said, it's kind of hilarious to see all the star reviewers being so mad at Intel, when all they did was release a couple of products at the top-end with practically no progress for a bit more money, and kind of really good products at the mid-range for an attractive price.

It's the year(s) of the strange: a time when even reviewers themselves are getting scolded on social media for their content about (un)available products. Consumers are frustrated, reviewers end up frustrated too. Expensive and/or low availability products are met with immediate rejection, double rejection if they also happen to occasionally lose against prior gen flagships.

 

[coercitiv]

we dont know if ring or mesh or what between the small and large cores

We kinda' know, it's the precious ring. The small cores are connected in clusters of 4 to the bus, allowing ADL to stay within 10 ring stops. (we don;t ahve a confirmed source for this, but it does fit perfectly and makes a lot more sense than mesh for consumer)

 

[jpiniero]

We kinda' know, it's the precious ring. The small cores are connected in clusters of 4 to the bus, allowing ADL to stay within 10 ring stops. (we don;t ahve a confirmed source for this, but it does fit perfectly and makes a lot more sense than mesh for consumer)

I agree but Alder/Raptor is probally the end of the line for the ring.

 

[mikk]

69fps vs 86fps,
@IntelGraphics
UHD 750 driver 9200 vs 9316. Half Life 2, 1080p High

I'm noticing *up* to 10% better performance in certain scenarios with the new @IntelGraphics
UHD 750 driver, however most games I've tested thus far show similar performance as the press driver.

https://twitter.com/x/status/1378643802672439296

Some good gains it seems in some scenarios, although even the 9316 driver is quite old (dated mid February), they have just added the device RKL-S ID into the old 9316 driver. Intel is really slow with driver updates lately.

This isn't what I said. What I said was that since there are many more games which are unsupported on Xe than for which Intel has official support, it's anybody's guess as to how a wide spectrum of games would perform(if some of them run at all). It is obvious that some of them will not do as well as the games that are on Intel's list, which is why the +50% figure isn't a realistic indicator of UHD 750's performance over UHD 630.

You are denying all the tests and you are claiming that the games who are not listed run 20% slower, you have to prove it! And of course this is what you are literally saying, supported games will get a 20% boost, otherwise you comment doesn't make any sense. You failed to understand that all games are supported per se over the supported Xe 3d APIs.

 

[JoeRambo]

Intel is really slow with driver updates lately.

Driver team is most likely mobilized to help with dGPU drivers, no resources to spare on side show that is iGPU atm probably. Good thing is that it is same architecture, so game fixes/optimizations will trickle back to iGPU drivers.

 

[DrMrLordX]

If you're old enough to remember it's like when the old champ fought the new one... Tyson vs Spinks back in '85.

Thought it was '88?

PPV subscribers were angry that they didn't get much of a fight.

 

[CakeMonster]

Back when Intel improved both IPC and IGP every year until Skylake, I was quite optimistic about where the IGP was heading. This is disappointing in that regard, as it would need to catch up more to even stay on that trajectory. Hopefully AL and 10nm fixes that, but I'm not holding my breath.

 

[Hulk]

Thought it was '88?

PPV subscribers were angry that they didn't get much of a fight.

Yes you're right. For some reason I thought it was '85. I should have looked it up before posting it. I got the right decade anyway

I was one of those PPV subscribers, or more precisely my father was. It seemed like Michael Spinks didn't want to fight and was beat before he entered the ring. He looked like a deer in the headlights. I think it was just a payday for him. He was 31-0 entering the ring, got KO'd and never fought again even though he was only 31 years old. I guess he had enough.

 

[DrMrLordX]

Back when Intel improved both IPC and IGP every year until Skylake, I was quite optimistic about where the IGP was heading. This is disappointing in that regard, as it would need to catch up more to even stay on that trajectory. Hopefully AL and 10nm fixes that, but I'm not holding my breath.

Honestly we just need Intel to release Alder Lake at all, preferably on an aggressive schedule. Intel has yet to prove they can produce a legit desktop-oriented chip on 10nm anything. 8c TigerLake-H is not coming to desktop it seems, so . . . it's Alder Lake or nothing. Allegedly Intel has shipped millions of 4c Tiger Lake-U. They need to do the same with Alder Lake-S.

 

[tamz_msc]

You are denying all the tests and you are claiming that the games who are not listed run 20% slower, you have to prove it! And of course this is what you are literally saying, supported games will get a 20% boost, otherwise you comment doesn't make any sense. You failed to understand that all games are supported per se over the supported Xe 3d APIs.

OK so this was my experience running Batman Arkham Knight - it would not detect the Intel drivers at all, and the game runs using Microsoft's Basic Display Driver. I contacted Intel support and this was their response:



You happy now?

 

[TheGiant]

Honestly we just need Intel to release Alder Lake at all, preferably on an aggressive schedule. Intel has yet to prove they can produce a legit desktop-oriented chip on 10nm anything. 8c TigerLake-H is not coming to desktop it seems, so . . . it's Alder Lake or nothing. Allegedly Intel has shipped millions of 4c Tiger Lake-U. They need to do the same with Alder Lake-S.

IMO Intel can now release TGL-H when rocket lake is out and soon icelake-sp
both rcl-s and icl-sp are a little meh and serve as equalizers
now desktop, laptop and server have the same IPC

my prediction is both alder lake and sapphire rapids will come when its needed =sales decline

until then Intel needs time to switch from 14nm to full throttle 10nm +o maybe TSMC

 

[DrMrLordX]

IMO Intel can now release TGL-H when rocket lake is out and soon icelake-sp

If they can produce enough viable dice. It would still undercut their entire Rocket Lake inventory if they produced a Tiger Lake-S based on Tiger Lake-H 8c.

my prediction is both alder lake and sapphire rapids will come when its needed =sales decline

Right now Intel's buoyed by heavy demand. But things may get ugly for Cascade Lake-SP; in fact, it they may already have.

 

[Dayman1225]

Allegedly Intel has shipped millions of 4c Tiger Lake-U. They need to do the same with Alder Lake-S.

Yeah official number is 30 million TGL chips shipped to date. 150+ design wins and expecting 400 TGL systems throughout this year (not sure what is included in that statement).

 

[DrMrLordX]

@Dayman1225

Interseting. I had heard only 4 million, though if Intel can ill afford to lie about having shipped 30 million. If so then at least they're executing properly on 4c 10SF dice.

 

[cortexa99]

OK so this was my experience running Batman Arkham Knight - it would not detect the Intel drivers at all, and the game runs using Microsoft's Basic Display Driver.

The most worrisome thing that I had considered before finally happen: the compatibility of Xe graphic might be the most serious problem as it being an entry-level perf IGP.

Performance wise it doesn't matter cuz IGP won't competent enough against discrete card, but if IGP could run smooth in old games it would be a good selling point.

And now driver/compatibility issues of old games kicks in would cause the Xe graphic hardly make any sense in game area......

Yeah we still have to wait RKL-Xe or latter generation like ADL-Xe to solve this problem, but seeing what's happening now and consider Intel's ability in this area I wouldn't put to much hope into it......

 

[Det0x]

Tom's has an interesting face off between the 11900K and 5900X.

If you're old enough to remember it's like when the old champ fought the new one... Tyson vs Spinks back in '85.

AMD Ryzen 9 5900X vs Intel Core i9-11900K: Rocket Lake and Ryzen 5000 CPU Face Off

Intel's 14nm Rocket Lake squares off with AMD's 7nm Ryzen 5000

www.tomshardware.com

Ah good old toms Hardware strikes again
They always manage to have such good and above average scores for AMD products
But good job getting the 11900k to 5.2 allcore.

 

[Dayman1225]

@Dayman1225

Interseting. I had heard only 4 million, though if Intel can ill afford to lie about having shipped 30 million. If so then at least they're executing properly on 4c 10SF dice.

Yeah Pat said “just under 30 million” in the video. I don’t think volume for TGL U/H will be an issue - ICL SP which is launching tomorrow could be another story though - but intel said with their early ship program 100K+ ICL SP units already shipped - high hopes that ICL is volume mover in server.

 

[dullard]

Intel finally released the first Rocket Lake graphics drivers this morning.

Intel Rocket Lake Graphics Driver Now Officially Available

It has always been there.

www.tomshardware.com

That might help some people in this thread.

 

[lobz]

OK so this was my experience running Batman Arkham Knight - it would not detect the Intel drivers at all, and the game runs using Microsoft's Basic Display Driver. I contacted Intel support and this was their response:

View attachment 42553

You happy now?

No. You have to prove it by blood.

 

[moonbogg]

When the i7 was about to be released, people speculated it would be bad at gaming. Hopefully we get surprised like that again.

 

[itsmydamnation]

OK so this was my experience running Batman Arkham Knight - it would not detect the Intel drivers at all, and the game runs using Microsoft's Basic Display Driver. I contacted Intel support and this was their response:

View attachment 42553

You happy now?

all i can think of is the end of this song.....lol

 

[lobz]

When the i7 was about to be released, people speculated it would be bad at gaming. Hopefully we get surprised like that again.

That's not how I remember this happening. I recall that the local GFA (Geekbench Fetish Association) was pretty adamant that RKL would be this gaming champion riding on the back of some absurd sounding IPC gains. So I'd say that the surprise was more like negative, and that came with the first real i7 "leaks".