*Official Benchmark Thread* - Post it here or it didn't happen :D

[Mr. Fox]

4 hours ago, electrosoft said:

I'm going to FAFO and give the Z890 Unify-X a whirl especially at $399.99 on Amazon. Worst case scenario, back it goes! I liked the MSI BIOS, but can readily admit I still find Asus BIOS to be my favorite especially now that EVGA no longer makes motherboards. 😞

 

I could see why they exited the GPU market, but I still don't understand why they also did away with their excellent motherboard division.  😠

Yeah it is a real bummer. I am puzzled as well. Maybe they didn't make enough on motherboard sales for it to be sustainable. I loved their hardware and firmware.

 

That will be a nice case upgrade for the wife. Antec makes good stuff, especially their cases.

 

I enjoyed the Intel Unify X, but the firmware kept me from loving the X870E Carbon. It was not terrible by any means, it just never rose to the level of seeming awesome... good versus great.

 

[electrosoft]

6 hours ago, electrosoft said:

In classic Asus fashion, new BIOS drops for the Z890-A Strix and now it won't boot G4 9200 at all.... 🤣 Going to retest 8000->8400+ but I'm not expecting any miracles.

 

 

 

So with the new BIOS and further tuning and goosing, I've encountered a few new things.

 

Vengeance 9200 sticks wouldn't boot with XMP timings because Corsair was Pushing Dram Write Recovery Time to 138 for some reason. That one, single XMP setting was shutting down the Strix every time. Took a little to track that one down.

 

This 265k gives up the ghost at 8400.

The Z890-A Strix gives up the ghost at 8800.

 

I can't do G2 8400 fail

I can do G4 8400 pass

I can do G4 8667 pass

I can't do G4 8800 fail

 

So the Strix-A is doing better than expected and it did fail HUB's 8800 check during its review so no shocker. May come down to binning boards at this lower tier. This is the entry level Strix board.

 

That Unify-X I ordered won't ship till next week so that gives me a few more days to at least see if I can get the board to 8800 which is my return/keep threshold.

 

If I can stabilize 8800 I'll keep it. If not, back it goes.

 

I'm thinking this 265k was an open box return to MC because someone tested the IMC and went "garbage!" and returned it. I'll probably go ahead and put some feelers on the OCN and/or grab at least 1-2 more 265k's to bin.

 

But the testing continues....

 

 

 

 

 

[Mr. Fox]

A test BIOS (9923) dropped from ASUS for AM5 and it seems like an improvement. I cannot get 8400 fully stable, but I think that might be a CPU/IMC limitation. This is my weakest CPU. I may try my better 9950X or my 4585PX, but there is almost no measurable difference in performance between 8200 and 8400, so no biggie either way.

 

These Kingbank XMP/EXPO 8400 sticks that @gupsterg turned me on to have been great. They were dirt cheap when I bought them on Amazon back in the "good old days" (not long ago) before the tech world went off the rails with the crazy AI nonsense. I've never seen so many stupid people that would otherwise be viewed as intelligent. Now I wish I hadn't sold that second 8400 kit. Kingbank no longer sells memory on Amazon, and even if they did the price would suck now. 😑

 

[Mr. Fox]

@Rage Set it must be DIMM dying season. Check this out. Weird short circuit on this PC at the DIMM module.

 

 

[Talon]

Haven't bothered updating to latest Asus Apex BIOS yet, but this has been my daily Arrow Lake tune for around 6 months. I just laugh when I see 80+ns as "fast" Arrow Lake memory setups. This is on a clean install with shitty virtualization/hypervisor running on the system which hurts latency a tad. It absolutely wrecks gaming performance on Arrow Lake. In Cyberpunk 2077, it's almost a 17% penalty at 1080p high settings in the benchmark. 

[electrosoft]

10 hours ago, Mr. Fox said:

A test BIOS (9923) dropped from ASUS for AM5 and it seems like an improvement. I cannot get 8400 fully stable, but I think that might be a CPU/IMC limitation. This is my weakest CPU. I may try my better 9950X or my 4585PX, but there is almost no measurable difference in performance between 8200 and 8400, so no biggie either way.

 

These Kingbank XMP/EXPO 8400 sticks that @gupsterg turned me on to have been great. They were dirt cheap when I bought them on Amazon back in the "good old days" (not long ago) before the tech world went off the rails with the crazy AI nonsense. I've never seen so many stupid people that would otherwise be viewed as intelligent. Now I wish I hadn't sold that second 8400 kit. Kingbank no longer sells memory on Amazon, and even if they did the price would suck now. 😑

 

 

I'm still on 1805 from last month (newest) on my hero , but I went and checked and a new firmware update was available for connected devices for stability. Makes sense. Since the last bios update, my front USB-A ports are suddenly sporadic in whether devices connect or not. Check to see if there is a firmware update available for your board.

 

3 hours ago, Mr. Fox said:

@Rage Set it must be DIMM dying season. Check this out. Weird short circuit on this PC at the DIMM module.

 

 

 

Yeah, if you look back through some of his other videos, DIMMs are dying with a quickness recently. I actually enjoy his Fix or Flop series.

 

If anything, it shows you the power of having spare parts laying around....

 

1 hour ago, Talon said:

Haven't bothered updating to latest Asus Apex BIOS yet, but this has been my daily Arrow Lake tune for around 6 months. I just laugh when I see 80+ns as "fast" Arrow Lake memory setups. This is on a clean install with shitty virtualization/hypervisor running on the system which hurts latency a tad. It absolutely wrecks gaming performance on Arrow Lake. In Cyberpunk 2077, it's almost a 17% penalty at 1080p high settings in the benchmark. 

 

Arrowlake is new to me  and my favorite part is the beginning stages of a brand new platform and slowly exploring all the nuances over months. Just realized fabric potentially comes into play with OC and optimal freqency. Looks like ~8600 might be the sweet spot for dialing in timings in correlation to AL OC especially for gaming. The deep dive will continue over the next several months.

 

I may keep this Strix board if that's the case since it can do 8667 and just focus on binning chips around ~8800.

 

I did put up a feeler on OCN and of course got hit with the mailbox full of "message my friend at IWILLSCAMYOUHARD AT SCAMUGUD DOTZ LULZ" to report and ignore as always....

 

Did get 9200 back up and running in G4. I'm also realizing these "heat blankets" as @Mr. Fox calls them are coming into play on some of these stress test runs...

 

Good video showing speed tiers (but without super deep timings optimizations) on 14th and Arrowlake:

 

 

------------------------------------------------------------------------

 

Whenever @Papusan shows new Crocs on the forums (timestamped) 🤣:

 

 

[Mr. Fox]

15 minutes ago, electrosoft said:

Since the last bios update, my front USB-A ports are suddenly sporadic in whether devices connect or not. Check to see if there is a firmware update available for your board.

Hmm. Interesting. There is a firmware update for both the Apex and the Strix. In fact, it appears to be the exact same firmware update for both systems... same SHA-256 hash even.

Quote

ASUS 4242 and PD-GX Firmware Windows Update Tool

Version 4242: 250708, IT8857: 0206

5.83 MB 2025/12/12

SHA-256 :99DFED62167F0F5CD43522AF6611ACF9A5E6D8433996910B9FDC6B4A53BA8E4A

Updated firmware to ver. 250708 for ASM4242 and V0206 (GX) for IT8857 to improve compatibility with connected devices.

The PD_UPDATE_TOOL_V0206.exe flash completed successfully on the Apex, but the second ASMx4242 firmware update doesn't seem to do anything. Running it gives a momentary cursor circle icon that goes away immediately. I rebooted and it still does not seem to run. I wonder if the Readme.txt file has erroneous information about installing both updates. I will see what it does on the Strix.

 

I haven't had any major USB issues except for my webcam would intermittently and randomly throw a device descriptor failure error message on the Strix, but never on the Apex. Maybe this will fix that.