Intel Core Ultra 9 285k 'Arrow Lake' Discussion/News ("15th gen") on LGA-1851

[Rroff]

What you’re looking is out there. Where you’re going wrong is trying to attach terms like ‘full spectrum’ and ‘real world’ to what are fringe results and not representative of typical use cases.

The actual power consumption over a whole day of working in an application(s) is hardly fringe results :facepalm: it is literally what people do every day rather than canned benchmarks.

 

[jigger]

The actual power consumption over a whole day of working in an application(s) is hardly fringe results :facepalm: it is literally what people do every day rather than canned benchmarks.

What applications aren’t represented exactly?

I have never been unable to find results accurately representing power consumption.

 

[Rroff]

What applications aren’t represented exactly?

I have never been unable to find results accurately representing power consumption.

Read what I said again.

 

[Shaolin Chicken]

Just ran on OCCT stress test for an hour on the the Corsair RAM with no errors
Kind of kicking myself that I didn't mess around further with the GSKILL Ram, but I'm not used to messing with memory settings.
I've already put in an RMA request for the GSKILL ram but may hold onto it, and see if I can get it to boot later.

 

[jigger]

Read what I said again.

Or share the names of those mysterious everyday applications.

 

[RSR]

Just ran on OCCT stress test for an hour on the the Corsair RAM with no errors
Kind of kicking myself that I didn't mess around further with the GSKILL Ram, but I'm not used to messing with memory settings.
I've already put in an RMA request for the GSKILL ram but may hold onto it, and see if I can get it to boot later.

That good news is that the system is working, but it is not great on the G.Skill front.

I wonder if it has to do with the CKD driver, as I have seen a few updates around that recently with UEFI updates etc..

 

[Vimes]

I have no idea about any aspect of CU functionality with the memory. But I initially tried to get some Gskill Trident 7200 memory working with my Z790 Maximus ROG Hero board, well for a little while.
The SA voltage needed to be increased, as well as others, for that to be stable.
Instead, I went for lower bandwidth, IIRC 620OMhz, but with faster CAS and secondary timings.
That seems to work out ok, with my 14700k.

When you are able to boot into Windows I think it was CPUz that gave me the timings of the GSKILL memory that must be preset at various bandwidth settings.

I imagine that so much of what you can get is also dependant on the memory controller on the CPU...?

 

[Shaolin Chicken]

I have no idea about any aspect of CU functionality with the memory. But I initially tried to get some Gskill Trident 7200 memory working with my Z790 Maximus ROG Hero board, well for a little while.
The SA voltage needed to be increased, as well as others, for that to be stable.
Instead, I went for lower bandwidth, IIRC 620OMhz, but with faster CAS and secondary timings.
That seems to work out ok, with my 14700k.

When you are able to boot into Windows I think it was CPUz that gave me the timings of the GSKILL memory that must be preset at various bandwidth settings.

I imagine that so much of what you can get is also dependant on the memory controller on the CPU...?

Could be maybe my 265K is not cut out for 8200 speeds?

 

[Shaolin Chicken]

That good news is that the system is working, but it is not great on the G.Skill front.

I wonder if it has to do with the CKD driver, as I have seen a few updates around that recently with UEFI updates etc..

Thanks RSR, do you have a link please, so I can read a bit more into it?
Do you think it's worth hanging on to the GSKILL Cudimms?

 

[Vimes]

Could be maybe my 265K is not cut out for 8200 speeds?

RSR will, no doubt, have a much better understanding and knowledge than myself. But, from what I understand, getting decent QVL noted memory, a decent board, an ounce or two of understanding, and then the last part, an amount of luck on your side as to the memory controller of the CPU can all go to determine what you are able to get when going beyond what Intel set as their maximum supported.
Assuming, of course, that same understanding applies to the ArL as it did to their previous gen(s).

Not too long ago I remember having any form of stability at 8000, albeit RL, was considered exceptional.

I take it you do not have MemTest86 baked into your MB..? For me it is there under the Tool section and can be ran from within the BIOS to, at least, test memory settings from there, not needing to boot into Windows etc.

Best of luck.

 

[Rroff]

Or share the names of those mysterious everyday applications.

What mysterious every day applications? as before go back read what I actually said.

 

[Shaolin Chicken]

RSR will, no doubt, have a much better understanding and knowledge than myself. But, from what I understand, getting decent QVL noted memory, a decent board, an ounce or two of understanding, and then the last part, an amount of luck on your side as to the memory controller of the CPU can all go to determine what you are able to get when going beyond what Intel set as their maximum supported.
Assuming, of course, that same understanding applies to the ArL as it did to their previous gen(s).

Not too long ago I remember having any form of stability at 8000, albeit RL, was considered exceptional.

I take it you do not have MemTest86 baked into your MB..? For me it is there under the Tool section and can be ran from within the BIOS to, at least, test memory settings from there, not needing to boot into Windows etc.

Best of luck.

Thanks Vimes, no memtest, it will not help anyway as it will not post at 8200 speeds.

 

[jigger]

What mysterious every day applications? as before go back read what I actually said.

You said Intel CPUs use less power which is false.

 

[Rroff]

You said Intel CPUs use less power which is false.

What I said was, and this is completely factual if like RSR and myself you use energy monitoring over daily use:

This is the problem with a lot of reviews not showing the full spectrum especially how Intel CPUs are still often significantly more power efficient than AMD in idle through to moderate utilisation while AMD currently hold the advantage under heavy utilisation.

So the actual power consumption over a whole day can be a very different story depending on your usage.

 

[jigger]

What I said was, and this is completely factual if like RSR and myself you use energy monitoring over daily use:

You see, I do even at home. That’s why I’m now almost exclusively using AMD platforms.

 

[Vimes]

Thanks Vimes, no memtest, it will not help anyway as it will not post at 8200 speeds.

Ah, fair enough. I wasn't sure if it would allow you to get into the BIOS, rather than not POST at all.

If that is the case then, and not POST at all, how do you know that the memory is being set to 8200Mhz, rather than one or both sticks faulty..? Doesn't the MB, be default, set outside of XMP1, as an example, at a more conservative setting..?

 

[Rroff]

You see, I do even at home. That’s why I’m now almost exclusively using AMD platforms.

You'd know then that Intel platforms have the advantage at low CPU utilisation states while AMD's advantage is under heavy CPU utilisation states so for a simplified example if a system was sitting there idle for 55 minutes of an hour but compressing files for 5 minutes of the hour over a whole day the Intel system would be more energy efficient despite in a canned compression benchmark the AMD system would use less power.

The same for something like Unreal Engine - you don't sit there compiling a level constantly all day which is what the canned benchmark does, the actual CPU usage of the application varies as you do different tasks in it and over a few hours of usage the actual power consumption can be very different to what canned benchmarks suggest - the actual result varying on usage patterns.

 

[RSR]

Thanks RSR, do you have a link please, so I can read a bit more into it?
Do you think it's worth hanging on to the GSKILL Cudimms?

Sure, I usually scan over the following threads on OCN.

Overclocking "Arrow Lake" - 285K, 265k, 245k etc results, bins and discussion - https://www.overclock.net/threads/o...11860/page-445?post_id=29453741#post-29453741
*Official* Intel DDR5 OC and 24/7 daily Memory Stability Thread - https://www.overclock.net/threads/o...4772/page-1640?post_id=29453334#post-29453334
ASRock Z890 TAICHI OC FORMULA Thread - https://www.overclock.net/threads/a...12427/page-123?post_id=29452506#post-29452506
[OFFICIAL] Asus Strix/Maximus Z890 Owners Thread - https://www.overclock.net/threads/official-asus-strix-maximus-z890-owners-thread.1812501/page-52

There are a few other sites, but OCN is fairly active.

 

[jigger]

You'd know then that Intel platforms have the advantage at low CPU utilisation states while AMD's advantage is under heavy CPU utilisation states so for a simplified example if a system was sitting there idle for 55 minutes of an hour but compressing files for 5 minutes of the hour over a whole day the Intel system would be more energy efficient despite in a canned compression benchmark the AMD system would use less power.

Outside of turning a system on (windows using a few specific chips) and letting it sit idle in its eco state doing nothing for the hell of it. Even then there is no clear advantage between Intel and AMD.

 

[RSR]

Could be maybe my 265K is not cut out for 8200 speeds?

I don't see why it shouldn't run, as you have a 1 DPC board (2 DIMM), which should be good for 9K+. What I find interesting, though, is that your 8K kit doesn't have a CKD driver, and the 8200 kit does, which makes me question if it was a UEFI update.

I guess I quick way to tell is to bump the speed to 8200 on the Corsair set and increase the VDD/VDDQ a little and see if it boots.