Question Raptor Lake - Official Thread

Page 47 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Hulk

Diamond Member
Oct 9, 1999
4,191
1,975
136
Since we already have the first Raptor Lake leak I'm thinking it should have it's own thread.
What do we know so far?
From Anandtech's Intel Process Roadmap articles from July:

Built on Intel 7 with upgraded FinFET
10-15% PPW (performance-per-watt)
Last non-tiled consumer CPU as Meteor Lake will be tiled

I'm guessing this will be a minor update to ADL with just a few microarchitecture changes to the cores. The larger change will be the new process refinement allowing 8+16 at the top of the stack.

Will it work with current z690 motherboards? If yes then that could be a major selling point for people to move to ADL rather than wait.
 
  • Like
Reactions: vstar

Abwx

Lifer
Apr 2, 2011
10,847
3,297
136
We? Who's we? You decided to use computerbase.

What if we use TPU? We know what their config is. DDR4-3600 C14 on AMD. DDR5-6000 C36 on Intel Alder Lake.

And the score here is, 25813.


View attachment 65528

That s 1.5% difference with Computerbase for the 5950X, the difference is only 1.7% for ADL, as you can see RAM has quite low influence in Cinebench.

Raichu alleged score for the 5950X is 24116 for the comparison with RPL@250W and 24320 with RPL350W according to the percentages he displayed....
 

tamz_msc

Diamond Member
Jan 5, 2017
3,710
3,554
136
Problem with Geekbench is it doesn't test much that actually stress a high end desktop level CPU enough to matter. Who cares if any of the micro benchmarks that make up Geekbench is fast when you use a desktop where it wouldn't be noticeable at all in everyday use? You can notice your frame rate is 10% lower, you can't really notice if takes 100 nanoseconds more to open a PDF. A lot of the Geekbench benchmarks aren't all that useful and they are so small it barely matters. There is very few use cases on desktop where the CPU is actually stressed and they are very specific to the person using the computer.

You can say everyone does web browsing so it's the best benchmark but I'm pretty sure nobody can tell 1 cpu from another with a blind test just doing web browsing.
Gaming, compiling, rendering, transcoding, productivity, etc are pretty specific to people, this means people need to check benchmarks that matter to them. Geekbench doesn't really allows for that and just does their own weights of a whole bunch of things and gives a score.
This simply isn't true. Before my present laptop that has an i7-11370H, I used a desktop with an i7-3770. I can definitely tell the difference between the two when using a web browser. Stronger cores, bigger cache, faster RAM - they all add up.
 

Makaveli

Diamond Member
Feb 8, 2002
4,715
1,049
136
Problem with Geekbench is it doesn't test much that actually stress a high end desktop level CPU enough to matter. Who cares if any of the micro benchmarks that make up Geekbench is fast when you use a desktop where it wouldn't be noticeable at all in everyday use? You can notice your frame rate is 10% lower, you can't really notice if takes 100 nanoseconds more to open a PDF. A lot of the Geekbench benchmarks aren't all that useful and they are so small it barely matters. There is very few use cases on desktop where the CPU is actually stressed and they are very specific to the person using the computer.

You can say everyone does web browsing so it's the best benchmark but I'm pretty sure nobody can tell 1 cpu from another with a blind test just doing web browsing. Gaming, compiling, rendering, transcoding, productivity, etc are pretty specific to people, this means people need to check benchmarks that matter to them. Geekbench doesn't really allows for that and just does their own weights of a whole bunch of things and gives a score.

The obsession with synthetic benchmarks scores is indeed a problem, the workflow and what you do on your machine matters the most.
 
  • Like
Reactions: Kaluan

shady28

Platinum Member
Apr 11, 2004
2,520
397
126
That s 1.5% difference with Computerbase for the 5950X, the difference is only 1.7% for ADL, as you can see RAM has quite low influence in Cinebench.

It is? I see >10% on R23 multi going from DDR4-3600 C16 to DDR5-6000.

Do you see something different?

1659897189451.png
 

pakotlar

Senior member
Aug 22, 2003
731
187
116
35,690 CB23 MT (2290 ST) at 250W default PPT/PL2, 40,000 when PL2 unlimited (hots 350W).


My 5950X gets 28,000 at 190W PPT and 25,000 or so at 142W. The more efficient number is 25,000/142 = 176 points/W. The OC value is 28,000/190 = 147 points/W. The OC number is the most efficient OC I’ve been able to achieve, a naive overclock would do worse (I will get 29,000 with 220W).

Raptor Lake gets 36,000/250 = 144, just a bit worse than the 190W figure for my 5950X, and 20% behind the 5950X at stock. AMD claims 7950X is 25% more efficient at stock. My guess is that is 25% more performance at ISO power, not at the juiced 230W PPT they need to hit 5.7ghz in the 7950X. However, lets give AMD the benefit. That would mean that 7950X would hit 36,000 at 163W. I will be very surprised, pleasantly if thats the case. I think it will hit around 36,000 at 230W, because Raichu or greymon previously wrote that Raptor Lake is likely faster, and Zen 3 barely scales with power beyond stock.

Overall, Raptor Lake looks like its just a bit behind Zen 3 in MT efficiency, and a good deal faster. A good refresh. We’ll see what Zen 4 brings, could be a close fight, or a blowout in Zen 4’s favor. Still Intel is moving in the right direction.
 
Last edited:

tomalbd77

Junior Member
Aug 7, 2022
1
1
36
Since we already have the first Raptor Lake leak I'm thinking it should have it's own thread.
What do we know so far?
From Anandtech's Intel Process Roadmap articles from July:

Built on Intel 7 with upgraded FinFET
10-15% PPW (performance-per-watt)
Last non-tiled consumer CPU as Meteor Lake will be tiled

I'm guessing this will be a minor update to ADL with just a few microarchitecture changes to the cores. The larger change will be the new process refinement allowing 8+16 at the top of the stack.

Will it work with current z690 motherboards? If yes then that could be a major selling point for people to move to ADL rather than wait.
important this post
 
  • Haha
Reactions: lobz

Carfax83

Diamond Member
Nov 1, 2010
6,841
1,536
136
12900K crypto score with DDR5-6400 was 4274. 13900K is doing 6266 with unknown DDR5 speed. So RPL may do really well in memory sensitive applications.

I think in the thread he said it was running DDR5 6400 CL34.

I didn't know that the crypto sub test was sensitive to memory performance, but he also says that DDR4 would yield a higher score?

I would need to see some Aida64 comparisons to come to a firm conclusion, as I know for a fact that specifically targets memory performance.
 

pakotlar

Senior member
Aug 22, 2003
731
187
116
Computerbase tests are done at manufacturer s stock settings, including the RAM speed and latencies, so my point is that the numbers exhibited by Raichu for the 5950X are plain wrong.
25,000 is closer to the base 5950x result than 26,000. Early reviews had it at 24,500, which is what my mediocore 5950x bin gets with potato 3600mhz cl18 ram.
 

shady28

Platinum Member
Apr 11, 2004
2,520
397
126
This simply isn't true. Before my present laptop that has an i7-11370H, I used a desktop with an i7-3770. I can definitely tell the difference between the two when using a web browser. Stronger cores, bigger cache, faster RAM - they all add up.

You are exactly right, and what's more, many people work exclusively on browsers. Modern browsers are very much like little VMs. People use them to stream multiple live charts of stocks and bonds, video conference, for MS office functions, email, and far more. These are totally normal day to day use cases. If you have an IRA or 401K that you keep track of closely, you're likely using these functions and as you say - I can tell a huge difference in web responsiveness using an old 6300U laptop vs an 8365U laptop even on accessing mundane things like Cigna's site for healthcare.

Even on desktop, I have had plenty of instances where my 10850K will hit a single core at 100% on a web based application. AES encryption is part of that as well, as every site uses it.
 
  • Like
Reactions: igor_kavinski

Zucker2k

Golden Member
Feb 15, 2006
1,810
1,159
136
250W is a bit much. I have an Arctic Liquid Freezer II 240 AIO and it can keep up with about 190W continuous. It's certainly fine for short periods of 240W or so but we're talking like 10s.

I think a very good 360mm AIO is probably good to 225-250W, but past that you'll need custom loop. Most people with a 360 AIO will wind up overheating at 250W.
Raichu posted this reply:

 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
The thing with cinebench is it's more like a horsepower test than anything else.
It might not be repressentative of common use cases but it is representative of what a cpu will be doing when it's under full tilt, even other software.

It's all FP though, and it doesn't really tax the memory subsystem or even some cache subsystems of certain CPUs. It's not perfect. Some renderers don't behave like it. Some fp benchmarks make better use of AVX512 (or even AVX2).

Also everything has a cinebench run so you can compare against everyting else, although with how rapid the c23 release was I'd argue it's going against that now.

There are a few CPUs that weren't tested against R23 in benchmark articles. But otherwise R23 scores are now quite ubiquitous.

Again spoken out of ignorance:

Wow you're in a spicy mood. Has it occurred to you I've read that? Now, how many people publishing GB5 scores know all that? And have you seen the actual source code?

Why do so many of the MT tests not even make full use of the CPU? What the hell is it doing in there?

Do you trust Geeklabs?

It's almost like people forgot how CBR15 used a different version of Maxon's renderer than their commercial software (making it useless for performance profiling for a time). Just because you like a particular closed source blob for benchmarking doesn't mean its indicative of how a CPU will perform for you with real applications.

It's an okay representation of like vs. like performance within the same uarch, say Alder Lake vs. Raptor Lake, though we've already seen some divergent Raptor Lake ES scores so it's hard to tell what to expect from the commercial product.
 
  • Like
Reactions: ZGR

shady28

Platinum Member
Apr 11, 2004
2,520
397
126

IDK. Some are disagreeing with Raichu and I disagree as well.

This is from kitguru. He's using a 5950X OC, temps are above ambient which he is reporting as 22-25C. IIRC, Zen 3 throttles at 90C.

The lowest temp here allowed the highest power draw at 226W, temp is 59+(22 to 25) = 81-84C. This chart pretty much reflects my experience. 240 AIO is good for in the area of 175-200W, the best 360s are usually good for 225ish, maybe 250 if you open the case and let the radiator blow clean air. After that, you're in no man's land.

There are some exceptions. Like Arctic makes a 420 AIO.

These are mostly 360mm AIOs at the top, and some tower air coolers bouncing off 90C at ~210W.

1659922225442.png
 

Markfw

Moderator Emeritus, Elite Member
May 16, 2002
25,482
14,434
136
It's all FP though, and it doesn't really tax the memory subsystem or even some cache subsystems of certain CPUs. It's not perfect. Some renderers don't behave like it. Some fp benchmarks make better use of AVX512 (or even AVX2).



There are a few CPUs that weren't tested against R23 in benchmark articles. But otherwise R23 scores are now quite ubiquitous.



Wow you're in a spicy mood. Has it occurred to you I've read that? Now, how many people publishing GB5 scores know all that? And have you seen the actual source code?

Why do so many of the MT tests not even make full use of the CPU? What the hell is it doing in there?

Do you trust Geeklabs?

It's almost like people forgot how CBR15 used a different version of Maxon's renderer than their commercial software (making it useless for performance profiling for a time). Just because you like a particular closed source blob for benchmarking doesn't mean its indicative of how a CPU will perform for you with real applications.

It's an okay representation of like vs. like performance within the same uarch, say Alder Lake vs. Raptor Lake, though we've already seen some divergent Raptor Lake ES scores so it's hard to tell what to expect from the commercial product.
Exactly. My 5950x scores 14k on that test, and I have checked, and it runs at up to 4.8 ghz, and no problem with heat with a 280 AIO. Why ?? I just don't trust that benchmark.
 
  • Like
Reactions: Kaluan and Drazick

pakotlar

Senior member
Aug 22, 2003
731
187
116
All of this arguing about whether 13900K at 350W is 67% faster than 5950x or “merely” 50% faster needs to stop, wait for reviews.

Both images are at 5ghz. And at 350W+, makes the 5950X look bad compared to Raptor Lake, though of course Zen 3 is way outside its comfort zone. Zen 4 will be better, maybe 25% at iso power, 43k at 350W?

 
Last edited: