Can Intel FIX Arrow Lake?

Is it possible for Intel to fix Arrow Lake before it's too late?
#shorts #intel #amd #arrowlake #zen5 #gaming #fix #cpu #bulldozer #ultra

MAJOR SPONSORS:

CDkeyoffer Cheap Windows & Gaming Keys:
25% software discount code: brokensilicon
3% OFF EVERYTHING discount code: dieshrink

Cash App Donations: $MooresLawisDead

Address Fan Mail to:
Moore's Law Is Dead,
PO Box 60632
Nashville, TN 37206

Broken Silicon [A PC Hardware Podcast] is available on ALL platforms.

All Music by SAHARA:

Broken Silicon Logo by Tom.
Broken Silicon burning graphic by Tom & Sulphurous:
2021 Video Intro by Tom & abstractrealism on the MLID Discord

LINKS!!!!!!:

[GrandmasterBBC]

That was very thoughtful of Intel to help AMD make their ryzen 9000 series chips not seem so lackluster.

[mattzun]

Unless Dell starts shipping desktops with AMD CPUs instead of Arrow Lake, Intel will consider Arrow Lake to be VERY successful.

Getting Arrow Lake out the door was important because the 13th/14th gen problems were bad enough that they might affect OEM sales/support costs.

In a world where Dell can sell Alienware PCs with absolutely terrible cooling for years, gaming performance is at best a trivial consideration.

[Shehryar]

Jays Two Cents used CUDIMs on Ultra 285, drastic improvement in performance.

[iLegionaire]

Intel cannot reverse the benchmarks favoring the 5800X3D over the Intel Core Ultra 9 285K.

Also Intel is releasing 3D-VCache in 2027 codenamed Aero Cache or Last Level Cache (Yes three years from now after Arrow Lake flopping). By then, AMD will have three years to improve X3D for Zen 6 and Zen 7, and further push their 3D-VCache technology.

Not looking good for Intel here.

[nipa]

ARL is DOA. They have to cut prices in half.

[thefreemarketeer]

The scheduler is likely the bulk of the issue, second to that is reviewers not using the proper, high speed ram which offsets the latency penalty of the tile based design.

Plenty of evidence of scheduler problems; over clocking E cores yields massive performance uplifts in gaming, a clear indication the scheduler is incorrectly utilizing the E cores in gaming workloads

[kamm]

It's already done, they're just adding fuel to 🔥

[LivedWater]

Yea sucks how they messed up so bad here. No way someone is going to be uninformed enough to actually buy these CPUS just for their socket and longevity... they need some actual worthy performance here and the 285k is worse than the 7700x. Probably the worst hardware optimization blunder in history, Intel is such a big company and they still blew it. Definitely still siding with AMD here

[oltorftheviking]

I love how this guys reasoning is “it’s at 5.4 ghz so there’s no more performance to be gained” 😂😂

[lairbox]

Unless there's a massive bug in the windows scheduler... 11 can surprise.

[Xenshen]

If they had been realistic with what it was and everything was 50-100 cheaper it would be great. I think this is just a stepping stone like Ryzen 1 and 2. It will get better.

[Dante_S_Turbo]

We know there's some windows taskmanager issues now that we have had some o.c.'ed examples. Overclocking the P cores drops performance, overclocking E cores increases gaming performance, games shouldn't be using E cores. Cudimm ram improves gaming performance just existing. in some games by 20%

[TheHangarHobbit]

Its actually WORSE than Bulldozer because say what you want about the FX line but you could get FX super cheap compared to Intel's pricing whereas today Intel is trying to charge more for chips that are worse than AMD.

[RobBCactive]

Brian's Tech Yes City idea of large cache doesn't work for Intel monolithic chips they just don't need it to scale performance at high frequency. Ryzen relies on cache more because going from CCD through IOD to DRAM adds latency, while Intel kept memory controllers closer.
HUB tested increased cache size and found limited gains, unlike that seen going from i3 to i7, the i7 to i9 cache increase just didn't help much.
The Admantine project was shelved, apparently on cost grounds which could be poor price/performance, while the CU-DIMM memory does seem to reduce ADL memory latency now patches are being tested.

[huynhvuquochuyofficial]

They use a diffrence architecture so they should update new bios firmwares, new drivers for windows. It's will take long time to optimize.

[Felale]

The gains are with overclocking and CUDIMM. How is techtube still not aware of this? Jay is the only one who has posted this.

[lokeung]

Will there be some kind of latency error like Zen5 in the micro code that can change by the BIOS update?

[nathangamble]

Arrow Lake already has way more cache than standard Zen 5. It's extremely unlikely that adding even more cache would help Arrow Lake's gaming performance significantly.
Evidence indicates that Arrow Lake's problems are caused by interconnect/ring latency. More cache can reduce the impact of the latency between the CPU cores and memory controller, but also increases the load on the ringbus and may require a reduction in bus frequency, which could result in a latency increase that outweighs any benefits.

[Demonoid]

I'm wondering if the "filler" chiplet is reserved for something similar to X3D. I've been wondering if they'd ever bring back their XE chips like the 9980XE and 10980XE. Or something similar to it. There are the rumors floating about of a potential Nova lake CPU with 16p cores and 32 E cores. 52 Total threads/cores certainly sounds like something extreme..

I'll be interested to see how Arrow lake is performing after some bios/microcode updates and perhaps some Windows updates. It's just kinda strange that it has pretty good performance in synthetic benchmarks, and productivity but just falls flat in a lot of gaming tests.

The Z890 motherboards have some really nice features and options too.. Just a bit of a let down.

[derek]

TYC never knows what he talking about.AMD seems to be getting good gains, when Intel is just going backwards. TYC needs to stop putting his 2 cents in because he does not make sense.