- Mar 3, 2017
- 1,777
- 6,791
- 136
They could mark that on the slide, like 3 (effective) with explanation at the end of the slide deck. I might be nitpicking but I think the quality of AMD press materials is getting worse instead of getting better. Actually that was 1 FP512 load per cycle and 0.5 store per cycle. But still the execution pipes won't go to waste if you do something more than shuffle memory around and keep data in the registers.
mostwanted002
Member
Intel Core Ultra 9 "Arrow Lake" QS CPU tested at 250W: 4% faster than 14900K on average - VideoCardz.com
Intel Core Ultra 9 QS tested Wake up enthusiasts, there’s a new leak from Jaykihn. This leaker has provided numerous details on Intel’s upcoming hardware, including Arrow Lake, Lunar Lake, and more. The information appears very reliable and has already been partially confirmed. The latest update...
videocardz.com
It's really joever for intel.. is it??
I really don't want to talk about Intel products in this thread.Intel Core Ultra 9 "Arrow Lake" QS CPU tested at 250W: 4% faster than 14900K on average - VideoCardz.com
Intel Core Ultra 9 QS tested Wake up enthusiasts, there’s a new leak from Jaykihn. This leaker has provided numerous details on Intel’s upcoming hardware, including Arrow Lake, Lunar Lake, and more. The information appears very reliable and has already been partially confirmed. The latest update...
It's really joever for intel.. is it??
In any case I say both x86-64 vendors have been caught slacking with desktop.
Last edited:
Last minute delay incoming? Wouldn't look too bad considering intel are cooking themselves, literally.
inf64
Diamond Member
- Mar 11, 2011
- 3,884
- 4,692
- 136
The CPUs are on sale, I doubt there will be any delays. Supply might be an issue though, supposedly very little 9950x samples were sent to reviewers.
Hail The Brain Slug
Diamond Member
- Oct 10, 2005
- 3,881
- 3,311
- 146
Really want to see the efficiency of both, but Granite as well as ARL-S seem to be a disappointment in terms of performance. Don’t care who wins by a tiny margin. Well, the silver lining for AMD is that they’re on N4P instead of N3, at least…Intel Core Ultra 9 "Arrow Lake" QS CPU tested at 250W: 4% faster than 14900K on average - VideoCardz.com
Intel Core Ultra 9 QS tested Wake up enthusiasts, there’s a new leak from Jaykihn. This leaker has provided numerous details on Intel’s upcoming hardware, including Arrow Lake, Lunar Lake, and more. The information appears very reliable and has already been partially confirmed. The latest update...
It's really joever for intel.. is it??
inf64
Diamond Member
- Mar 11, 2011
- 3,884
- 4,692
- 136
Sorry, I meant for preorder at certain stores:
Funtech
Specializirana spletna trgovina s profesionalno računalniško opremo, gamersko opremo in zabavno tehnologijo.
www.funtech.si
https://geizhals.eu/amd-ryzen-9-9900x-100-100000662wof-a3202550.html
Hail The Brain Slug
Diamond Member
- Oct 10, 2005
- 3,881
- 3,311
- 146
It hasn't been unheard of for some webshops to assume a price and start taking pre-orders before a hard price and release date is available. I don't think this means anything for either.
FlameTail
Diamond Member
- Dec 15, 2021
- 4,384
- 2,761
- 106
Charlie's got an article:
www.semiaccurate.com
AMD’s Zen 5 is a missed opportunity in messaging
Last week AMD did their ‘Tech Day’ and it was anything but.
www.semiaccurate.com
Charlie's got an article:
AMD’s Zen 5 is a missed opportunity in messaging
Last week AMD did their ‘Tech Day’ and it was anything but.
I think that's good advice.Charlie said:
But it is funny how much AMD bends over backward to accommodate Microsoft (Pluton, bigger NPU) and keep getting shafted for it (not used in any current Surface products, missing CoPilot PC features despite being adequate).
Charlie's got an article:
AMD’s Zen 5 is a missed opportunity in messaging
Last week AMD did their ‘Tech Day’ and it was anything but.
No real new information in that article.
Charlie just copied AMD's slides and is complaining and whining excessively, which is of course his style of writing and sometimes quite nice to read.
Nevertheless he should have been reporting and commenting on the new features of zen 5. Regardless how bad AMD's tech day supposedly was.
Otherwise why write an article in the first place ?
How do these preliminary scores compare with Zen 5?
It doesn't look to me like 15% to 25% IPC uplift MLID has been talking about:
It doesn't look to me like 15% to 25% IPC uplift MLID has been talking about:
In CB r23, that QS at 253W gets beaten by 9950X ES between 160 to 200W.
Pat better get to church fast and pray for a miracle.
It’s a tough comparison. If we’re considering a water-cooled 9950x with PBO+CO, vs Intel 285K stock; it isn’t that much of a surprise to beat it.
Let’s see like for like, and then the scores will speak for themselves. I think it’ll be very close for pure performance.
Edit: If it’s not even, or Intel slightly ahead then they would’ve lost the MT advantage they had (with the caveat of instability/degradation, no small thing). Looks like ST is neck and neck as well. Now if Intel is a lot better on efficiency due to those E cores, maybe it won’t look so bad.
Last edited:
Here is an article where Mike Clark apparently confirms the differing AVX-512 implementation of Strix's cores (small and big cores).
https://www.techpowerup.com/324873/...d-zen-5c-cpu-cores-have-256-bit-fpu-datapaths
AMD says that the 9950x scores ~42k in Cinebench r23 MT at stock. That's the number I would go with, though the power consumption isn't known. It can use up to 230W, but there's reason to believe it doesn't actually use quite that much.
adroc_thurston
Diamond Member
- Jul 2, 2023
- 7,065
- 9,807
- 106
Newsflash: it isn't.
LNC is 14% off the RWC which is a real, real bad baseline (itsa regression off RPC despite the L2 bump).
What do you mean with regression? It's not slower, right? Or do you think about performance per area or something?
Saylick
Diamond Member
- Sep 10, 2012
- 4,033
- 9,454
- 136
They mean that Lion Cove (LNC) is advertised as being +14% IPC over Redwood Cove (RWC), which is used in Meteor Lake (MTL). However, RWC is slower than Raptor Cove (RPC), which is used in Raptor Lake. Desktop Arrow Lake (ARL-S) uses LNC P-cores and is the successor to Raptor Lake. Therefore, the expected IPC uplift from Raptor Lake's P-cores to Arrow Lake's P-cores is likely less than 14% IPC uplift.
Josh128
Golden Member
- Oct 14, 2022
- 1,318
- 1,983
- 106
So back to Zen 5. Apparently 90C on Zen 5 does not equal 90C on Zen 4 due to different thermocouple/RTD placement.
We also know Zen 4 hits 220W-230W stock on good cooling in full MT workloads but Zen 5 seems to top out at ~190ish in its stock mode, despite displaying cooler core temps than Zen 4. Its a strange thing-- I'd be very curious if 200W PPT Zen 4 = 200W PPT Zen 5 if you actually measured power draw from the wall. They use the same mobos, so it would be a valid comparison. Just seems weird that they wouldnt choose to make Zen 5 hit max PPT out of the box to get that sweet ~44-45K number.
We also know Zen 4 hits 220W-230W stock on good cooling in full MT workloads but Zen 5 seems to top out at ~190ish in its stock mode, despite displaying cooler core temps than Zen 4. Its a strange thing-- I'd be very curious if 200W PPT Zen 4 = 200W PPT Zen 5 if you actually measured power draw from the wall. They use the same mobos, so it would be a valid comparison. Just seems weird that they wouldnt choose to make Zen 5 hit max PPT out of the box to get that sweet ~44-45K number.
The memory latency from on-chip memory controllers to crossing from chiplet to chiplet seems to be contributing (negatively), the same as with MTL.
Intel has a quite a bit of total cache to compensate for it. 76 MB L2+L3. That's between 40 MB for plain 9700 and 104 MB for x3d
Saylick
Diamond Member
- Sep 10, 2012
- 4,033
- 9,454
- 136
Yeah, according to Zen Daddy Mike Clark, they put the temperature sensor closer to the hotspots so it's more accurate in Zen 5. In Zen 4, the temp sensor has a larger built-in margin to account for the lower accuracy in its placement, meaning the sensor would artificially report a higher temp because it wasn't exactly in the ideal spot (for safety reasons). If the placement is more ideal, that artificial buffer can be reduced and in theory the chip can push itself harder.
Why should he report on the new features of Zen 5? You can get that everywhere on the internet, that's not what people come to Charlie's site for. What you want is for them to release that information, give people like him time to digest it and come up with a list of questions, then provide enough Q&A time for him and other members of the press to ask and have answered their questions about the information.
It sounds like they barely allotted any time for Q&A, and to the extent they did they turned every question into something they could answer with "AI". Obviously they were told to do so by some marketing idiot up on high.
That's the kind of thing you read him for - to understand what is happening behind the scenes at a company. So why does AMD appear unwilling to talk about Zen 5 other than in context of "AI AI AI"?
Pointing out the Pluton thing is good too, as AMD apparently LEFT IT OUT OF THE BLOCK DIAGRAM until Charlie (and perhaps others, to be fair) called them out on it. Why do they want to hide that? Could it be because they know while it is something Microsoft wants, it is not something any of their customers want?
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)
- Started by DisEnchantment
- Replies: 25K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes + WCL Discussion Threads
- Started by Tigerick
- Replies: 22K
-
News NVIDIA and Intel to Develop AI Infrastructure and Personal Computing Products- Started by poke01
- Replies: 384
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
Facebook
Twitter