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

[Mr. Fox]

9 minutes ago, Mr. Fox said:

If it is SPD corruption and not a hardware failure the firmware can be copied from the good module to the corrupted one. If I can find the link to where I saw how to do it (overclock.net) I will post it. You have to have a ROG Maximum motherboard or the tool will not work. Unless something has changed it will not work with a Strix, Prime or TUF motherboard.

@Rage Set not only OpenRGB but also Corsair iCUE has been known to corrupt the SPD. If that is the problem and not a hardware failure, here is the link that I discovered and the method I used to fix the corrupted SPD more than once.

https://www.overclock.net/threads/corrupted-ddr5-spd-fix-tutorial.1795935/

 

I see numerous links about Corsair iCUE doing the same thing as OpenRGB. You can also fix it another way (which I have not tried).

https://linustechtips.com/topic/1505572-am4-x570-system-no-longer-posts-with-4-sticks-of-memory-installed/?do=findComment&comment=15930927

[Rage Set]

1 hour ago, Mr. Fox said:

It corrupted the stick in the first slot and the system would boot if I moved the good one to the first slot and let the memory train, then turn it off and put the corrupted stick in the second slot. It would not boot at all with the corrupted stick in the first slot. In Windows, the corrupted stick showed everything was erased. This actually happened like 3 times. The first time I did not know what happened and the second and third times I fixed it.

 

I got the system to boot with both sticks, but while the bad stick is powered (raindow RGB flowing), it is not being detected. I am going to assume the Asus OCTools are exclusive to the Apex boards. As I can't find them and my AI Suite - ASUS Dual Intelligent Processors app doesn't have anything related to OC'ing RAM. So I guess, RMA is the route I must take.

 

Thanks for helping bro!!!

[Rage Set]

5 minutes ago, Mr. Fox said:

@Rage Set not only OpenRGB but also Corsair iCUE has been known to corrupt the SPD. If that is the problem and not a hardware failure, here is the link that I discovered and the method I used to fix the corrupted SPD more than once.

https://www.overclock.net/threads/corrupted-ddr5-spd-fix-tutorial.1795935/

 

I see numerous links about Corsair iCUE doing the same thing as OpenRGB. You can also fix it another way (which I have not tried).

https://linustechtips.com/topic/1505572-am4-x570-system-no-longer-posts-with-4-sticks-of-memory-installed/?do=findComment&comment=15930927

 

LOL, I was literally reading the first link and saw your reply in that thread. 

[Mr. Fox]

Just now, Rage Set said:

 

I got the system to boot with both sticks, but while the bad stick is powered (raindow RGB flowing), it is not being detected. I am going to assume the Asus OCTools are exclusive to the Apex boards. As I can't find them and my AI Suite - ASUS Dual Intelligent Processors app doesn't have anything related to OC'ing RAM. So I guess, RMA is the route I must take.

 

Thanks for helping bro!!!

I downloaded the OCTool from HWBOT forum. Let me see if I can locate the link. You don't need an Apex. A Maximus board of any kind will work if memory serves me correctly.

 

When I stopped trying to control the memory RGB lighting and removed any and all software to control it (including Armory Crap) the problem stopped and never returned.

[Mr. Fox]

17 minutes ago, Rage Set said:

 

LOL, I was literally reading the first link and saw your reply in that thread. 

What kind of DDR5 motherboards do you have? I do not know if it will work on an AMD system.

 

The ASUS OCTools are not specific to ASUS, but some features only work with a limited set of boards where i2c bus access is available using software. I do not think ASUS authors the tools, but they are commonly used by people like Safedisk that work for/with ASUS.

 

The same OCTools allowed me to override the Intel ARC GPU power limits and voltage to overclock it further. (It has an ARC tool feature that worked similar to the EVGA Classified tool.) I discovered that in a Skatterbencher video.

 

Edit: @Rage Set I believe it is bundled in this OCPak archive.

I found a link to it in this thread: https://community.hwbot.org/topic/220910-asus-rog-maximus-apex-z790-apex-15/#comments

[electrosoft]

1 hour ago, electrosoft said:

 

And lastly, 265k is a potential issue....but we will see.

 

Testing so far with the Corsair 9200 sticks

 

G2 8000 pass

G2 8200 pass

G2 8400 fail (and I mean within the first minute)

G4 8400 pass 

 

8400 G4 pass is the same as with the Kingbank sticks except the KB sticks required a fan and 1.5v. These 9200 sticks only require 1.45v and no fan.

 

I'll goose the voltages on the IMC to see how much further I can push it in G2 but falling off the rails at G2 8400 isn't a great portend of things to come. 

 

But first I'm going to see how far the MB can go in G4 just to see how far it can go with the newest BIOS. If it can do at least 8800 I'll keep it. If not, back to Amazon it goes and most likely a Unify is in my future unless I can find a well priced and properly working Apex.

 

Correction. I wasn't paying attention (in a rush today since it is the wife and my official holiday shopping all day "fun"), I forgot when you switch gears it auto resets the memory frequency to auto so that G4 pass run was at 5600 (DERP!). This means most likely last time in G4 it was back to auto frequency too when testing the KB sticks and it passed.

 

So I set it to G4 to remove the CPU from the equation, set to 8400 and it failed again on both the KB and Corsair sticks. So that is a double failure at G4 on the KB 8200 sticks and Vengeance 9200 sticks when testing 8400 but 8200 passes.

 

So now it is pointing to the MB not the CPU.

 

I'll dive back in either very late tonight or tomorrow as now onward to Xmas shopping and dinner and more shopping....

 

 

[Mr. Fox]

18 minutes ago, Mr. Fox said:

Edit: @Rage Set I believe it is bundled in this OCPak archive.

I found a link to it in this thread: https://community.hwbot.org/topic/220910-asus-rog-maximus-apex-z790-apex-15/#comments

There might be a newer version in the opening post by Safedisk in this thread:

https://community.hwbot.org/topic/242375-rog-crosshair-x870e-heroapexextreme/

 

I will have to see if it works on my X870E Apex motherboard.

 

https://www.dropbox.com/scl/fi/94dljuo7lj4o0luen1hc1/OCPak_1506.zip?rlkey=sgozvyrcy88kxxhbo3fpwdcya&e=1&st=9zd14xd6&dl=0

[Rage Set]

17 minutes ago, Mr. Fox said:

There might be a newer version in the opening post by Safedisk in this thread:

https://community.hwbot.org/topic/242375-rog-crosshair-x870e-heroapexextreme/

 

I will have to see if it works on my X870E Apex motherboard.

 

https://www.dropbox.com/scl/fi/94dljuo7lj4o0luen1hc1/OCPak_1506.zip?rlkey=sgozvyrcy88kxxhbo3fpwdcya&e=1&st=9zd14xd6&dl=0

 

I am using my Asus Z690 Extreme board. The good news, the newer version works with my board and I was able to read and save a copy of the working stick's firmware. The bad news, is the software doesn't see the other stick. 

 

EDIT: Yes, I had iCue on this machine. So it is likely what caused the issue with the bad stick. 

[Mr. Fox]

6 minutes ago, Rage Set said:

 

I am using my Asus Z690 Extreme board. The good news, the newer version works with my board and I was able to read and save a copy of the working stick's firmware. The bad news, is the software doesn't see the other stick. 

 

EDIT: Yes, I had iCue on this machine. So it is likely what caused the issue with the bad stick. 

Try with the naked green sticks as well as the RGB. That's a bummer that it doesn't see the other stick. That suggests it could be a hardware failure. Did you enable SPD Write in the BIOS? If not, enable that and try again and maybe it will work.

[Rage Set]

21 minutes ago, Mr. Fox said:

Try with the naked green sticks as well as the RGB. That's a bummer that it doesn't see the other stick. That suggests it could be a hardware failure. Did you enable SPD Write in the BIOS? If not, enable that and try again and maybe it will work.

 

So this is what I found out thus far.

 

The green stick - is dead as a door nail. The PC won't boot with it, at ALL. Even with the known working stick in A2, trained and etc. As soon as I put it in any slot outside of A2, the mobo just clicks whenever I press power. 

 

The bad G.Skill stick will boot with the known working stick in A2 after it has trained. I don't think it is hardware related, but it is definitely corrupted by that damn iCue software. I tried changing around the speeds to see if it could be detected. I even lowered the trained stick to 2133, lol. Still can't be detected but obviously it is getting power by the slot. It's almost behaving like a dummy stick. 

[Mr. Fox]

23 minutes ago, Rage Set said:

 

So this is what I found out thus far.

 

The green stick - is dead as a door nail. The PC won't boot with it, at ALL. Even with the known working stick in A2, trained and etc. As soon as I put it in any slot outside of A2, the mobo just clicks whenever I press power. 

 

The bad G.Skill stick will boot with the known working stick in A2 after it has trained. I don't think it is hardware related, but it is definitely corrupted by that damn iCue software. I tried changing around the speeds to see if it could be detected. I even lowered the trained stick to 2133, lol. Still can't be detected but obviously it is getting power by the slot. It's almost behaving like a dummy stick. 

If you can determine the target slot number with the tool after saving the SPD from the good stick, select the target slot for the dead stick and see if the tool will allow you to write to it. When I had the SPD corruption happening from OpenRGB the stick seemed dead. CPU-Z could not "see" it and the tool could not read from it, and Thaiphoon Burner showed the module had no firmware on it. But, I was able to select the target slot annd write the SPD from the good stick to the corrupted one. Set the BIOS for SPD Write enabled (or SPD write disable set to "enabled" depending on how it is worded in the BIOS) has to be toggled. It may not work, but it would be worth a shot. Maybe you will get lucky.

[Rage Set]

9 minutes ago, Mr. Fox said:

If you can determine the target slot number with the tool after saving the SPD from the good stick, select the target slot for the dead stick and see if the tool will allow you to write to it. When I had the SPD corruption happening from OpenRGB the stick seemed dead. CPU-Z could not "see" it and the tool could not read from it, and Thaiphoon Burner showed the module had no firmware on it. But, I was able to select the target slot annd write the SPD from the good stick to the corrupted one. Set the BIOS for SPD Write enabled (or SPD write disable set to "enabled" depending on how it is worded in the BIOS) has to be toggled. It may not work, but it would be worth a shot. Maybe you will get lucky.

 

With the bad stick in and when the mobo tries to retrain, I get error Code 55. I am going to chaulk it up as bad and submit an RMA with G.Skill. I am going to spend a lot of money to make sure it gets to them, which means insuring it for $600. I am going to ship it via FedEx, so hopefully it doesn't go missing. 

 

Thanks for all of your help brother. It felt great getting into the minutiae of hardware troubleshooting like this again - I don't get a lot of that in my IT day job.

[Mr. Fox]

2 hours ago, Rage Set said:

 

With the bad stick in and when the mobo tries to retrain, I get error Code 55. I am going to chaulk it up as bad and submit an RMA with G.Skill. I am going to spend a lot of money to make sure it gets to them, which means insuring it for $600. I am going to ship it via FedEx, so hopefully it doesn't go missing. 

 

Thanks for all of your help brother. It felt great getting into the minutiae of hardware troubleshooting like this again - I don't get a lot of that in my IT day job.

You're welcome. I also enjoy that kind of thing. If/when you can figure something out, or explore something out of the ordinary trying to figure it out, it feels good.

[Reciever]

Got about half of the watercooling gear today, for all the hate that EK gets the WB certainly feels premium, fairly heavy too lol, also got the 1080mm rad as well so I can take my time routing cables and such 

[electrosoft]

Ugh, full day of Xmas shopping with the wife..... me at the end of the day.....

 

 

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

 

Time to pick another motherboard as the TG 8200 sticks, which I had running at 8400 no problem on Z790i Lightning and 14900KS also borked out in G2 and G4. So that's three sets of sticks (8200, 8400, 9200) giving up the ghost on the Z890 Strix-A in G2 and G4 at 8400.

 

Even boots at 9200 G4 and AIDA64 run with the Vengeance sticks, but passing TM5 is a whole different story...

 

This is when having other CPUs / MBs is a plus to start bouncing off each other. At least I have 3x sets of DDR5 in the kitty though so that's a blessing.

 

 

 

 

 

[Papusan]

 

This was/is inevitable... Those with sligth older gen cards/new low mid end saved by  the dram crisis. 

 

Game development is now affected by the RAM crisis

 

TL;DR: Rising DDR5 memory and storage costs, driven by the ongoing DRAM crisis and AI demand, are causing supply shortages and higher prices through 2026. This forces game developers like Larian Studios to optimize AAA titles for limited RAM and VRAM, impacting PC gaming performance and hardware upgrade cycles.

 

What does this mean for gaming, specifically high-profile AAA-style experiences with a focus on immersive and detailed visuals? Well, a fundamental shift in how game developers will approach making games for the "foreseeable future." And with that, game developers will be forced to optimize performance for systems with limited RAM and VRAM.

 

 

G.Skill And Micron Share Depressing Update On PC Memory Crisis

by Paul Lilly - Thu, Dec 18, 2025

Just in case anyone needs to hear it again (as if!), we're in the midst of a memory crisis that has seen the cost of DRAM catapult to unsettling levels. 

 

G-skill statement.

https://www.gskill.com/community/1502239313/1765875574/Why-has-G.SKILL-DRAM-memory-prices-increased-so-much-recently-since-2025-Q4

 

[Rage Set]

19 minutes ago, Papusan said:

 

This was/is inevitable... Those with sligth older gen cards/new low mid end saved by  the dram crisis. 

 

Game development is now affected by the RAM crisis

 

TL;DR: Rising DDR5 memory and storage costs, driven by the ongoing DRAM crisis and AI demand, are causing supply shortages and higher prices through 2026. This forces game developers like Larian Studios to optimize AAA titles for limited RAM and VRAM, impacting PC gaming performance and hardware upgrade cycles.

 

What does this mean for gaming, specifically high-profile AAA-style experiences with a focus on immersive and detailed visuals? Well, a fundamental shift in how game developers will approach making games for the "foreseeable future." And with that, game developers will be forced to optimize performance for systems with limited RAM and VRAM.

 

 

G.Skill And Micron Share Depressing Update On PC Memory Crisis

by Paul Lilly - Thu, Dec 18, 2025

Just in case anyone needs to hear it again (as if!), we're in the midst of a memory crisis that has seen the cost of DRAM catapult to unsettling levels. 

 

That's actually great news. It will force developers to optimize for all levels, instead of the high end. Don't get me wrong, this will make those that have invested thousands on high end GPUs a bit angry, but those people only account for about 10% of the market. 

 

The true adverse thing to come out of this RAM shortage is Nvidia and AMD will continue to push 8GB cards in their next generations. Or worse, the reintroduction of the 6GB or lower "midrange" GPUs....Lol the 6070 coming with 8GB and 6070 TI with 12GB above $750

[Mr. Fox]

1 hour ago, Papusan said:

 

This was/is inevitable... Those with sligth older gen cards/new low mid end saved by  the dram crisis. 

 

Game development is now affected by the RAM crisis

 

TL;DR: Rising DDR5 memory and storage costs, driven by the ongoing DRAM crisis and AI demand, are causing supply shortages and higher prices through 2026. This forces game developers like Larian Studios to optimize AAA titles for limited RAM and VRAM, impacting PC gaming performance and hardware upgrade cycles.

 

What does this mean for gaming, specifically high-profile AAA-style experiences with a focus on immersive and detailed visuals? Well, a fundamental shift in how game developers will approach making games for the "foreseeable future." And with that, game developers will be forced to optimize performance for systems with limited RAM and VRAM.

 

 

G.Skill And Micron Share Depressing Update On PC Memory Crisis

by Paul Lilly - Thu, Dec 18, 2025

Just in case anyone needs to hear it again (as if!), we're in the midst of a memory crisis that has seen the cost of DRAM catapult to unsettling levels. 

 

G-skill statement.

https://www.gskill.com/community/1502239313/1765875574/Why-has-G.SKILL-DRAM-memory-prices-increased-so-much-recently-since-2025-Q4

 

 

38 minutes ago, Rage Set said:

 

That's actually great news. It will force developers to optimize for all levels, instead of the high end. Don't get me wrong, this will make those that have invested thousands on high end GPUs a bit angry, but those people only account for about 10% of the market. 

 

The true adverse thing to come out of this RAM shortage is Nvidia and AMD will continue to push 8GB cards in their next generations. Or worse, the reintroduction of the 6GB or lower "midrange" GPUs....Lol the 6070 coming with 8GB and 6070 TI with 12GB above $750

The current situation sucks. However, there is always a silver lining to a dark cloud when you are forced to look for it.

 

This will likely benefit those selling spare used PC components. They (we) may be able to sell things used for the same amount or more than what was paid for them new. And, an additional benefit is that consumers will probably stop wasting money upgrading systems that still work fine for the tasks for which they are used.

 

If games are dumbed down or "optimized" to function well on less capable hardware, the notion that an upgrade is needed should also be dramatically reduced. In practical application, systems that were nearing the end of their service life due to performance limitations may get a new lease on life.

 

Another potential benefit is this may pressure the Redmond Reprobates to back off on their idiotic OS system requirements if their intent was to contribute to e-waste development by forcing people to buy newer computers. That may become more difficult for consumers and cause Winduhz 11 adoption to stagnate even further than it has.

 

And, another potential benefit... perhaps the best possible scenario of all... AI development will implode, end up in a state of ruin and self-destruct. The very thing causing these problems may burn itself out due to astronomical costs developing something that their intended victims will not be able to leverage due to not enough consumers having the resources needed to utilize the garbage they are hoping to capitalize on. The idea of spending gobs of money on something that will not generate enough revenue to offset costs will ultimately fail. This is what I am hoping happens.

 

 

[Talon]

10 hours ago, electrosoft said:

Time to pick another motherboard

 

APEX. 

[electrosoft]

2 hours ago, Papusan said:

 

This was/is inevitable... Those with sligth older gen cards/new low mid end saved by  the dram crisis. 

 

Game development is now affected by the RAM crisis

 

TL;DR: Rising DDR5 memory and storage costs, driven by the ongoing DRAM crisis and AI demand, are causing supply shortages and higher prices through 2026. This forces game developers like Larian Studios to optimize AAA titles for limited RAM and VRAM, impacting PC gaming performance and hardware upgrade cycles.

 

What does this mean for gaming, specifically high-profile AAA-style experiences with a focus on immersive and detailed visuals? Well, a fundamental shift in how game developers will approach making games for the "foreseeable future." And with that, game developers will be forced to optimize performance for systems with limited RAM and VRAM.

 

 

G.Skill And Micron Share Depressing Update On PC Memory Crisis

by Paul Lilly - Thu, Dec 18, 2025

Just in case anyone needs to hear it again (as if!), we're in the midst of a memory crisis that has seen the cost of DRAM catapult to unsettling levels. 

 

G-skill statement.

https://www.gskill.com/community/1502239313/1765875574/Why-has-G.SKILL-DRAM-memory-prices-increased-so-much-recently-since-2025-Q4

 

 

Like @Rage Set and @Mr. Fox said, this is a blessing and a curse. Actually forcing optimizations in software to compensate is a blessing. Luckily (unluckily?) the memory shortage hits VRAM, RAM and potentially storage.

 

If it only hit VRAM, nothing would really change. You just lower the resolution and asset requirements.

RAM? Well, now you're talking not only assets but asset management and actual game code

Storage? Well, not you have to be cognizant of your total install size which encompasses assets, asset management and game code.

 

I think when all is said and done, high end GPUs won't be hit as much because texture sizes are variable and selectable. Even now, most 6GB video cards can run just about every game out there if you lower the detail level unless they take a shared memory model approach and dump off into normal system ram to compensate for texture sizes.

 

As a precaution, I picked up a few Samsung 9100 Pro 1TB NVME's a bit ago and prices have already spiked from what I paid (~$130) to now $170 each.

 

I already had a few 2TB SSDs (990 Pro, 9100 Pro) and a boatload of laptop SSDs (512GB to 2TB) sitting in the ~7 laptops on my shelf. Some of these have to go, though. 

 

49 minutes ago, Talon said:

 

APEX. 

 

I feel like a mic drop was done right after saying that.... 🤣

 

It is most likely going to be a UNIFY or APEX.

 

Do you plan on using an Apex for some 270k/290k binning?

 

Maybe I'll put out a WTB on the OCN forums for a tested and certified 9000 G2 Unify or Apex.

 

 

 

 

[Papusan]

 

It seems Nvidia have no problem serve the users with loads of vram if they can pay for it. 

 

NVIDIA Announces General Availability of RTX PRO 5000 72 GB GPU

Press Release 

The NVIDIA RTX PRO 5000 72 GB Blackwell GPU is now generally available, bringing robust agentic and generative AI capabilities powered by the NVIDIA Blackwell architecture to more desktops and professionals across the world.

 

With the flexibility to choose between this 72 GB variant and the existing NVIDIA RTX PRO 5000 48 GB model, AI developers can right-size their systems for a wider range of budgets and project requirements.

 

The forecast forwards. 3x price hike in 2025, aint enough...
DRAM memory will skyrocket in 2026 with increases of up to 80%, while NAND Flash will climb by 50%. Goldman Sachs

 

A particularly relevant point in the report is that Goldman Sachs does not identify HBM as the primary driver of price increases , but rather as a secondary factor. The bulk of the inflationary pressure comes from conventional memory, which is the most sensitive to capacity cuts and has the most direct impact on the consumer market. This implies that the price increases will not be limited to data centers, but will drastically affect PCs, laptops, smartphones, and home storage next year.

 

On 12/1/2025 at 5:27 PM, Mr. Fox said:

Welp. I guess I am done with EA forever now. Those scurvy bastards deployed the Javelin Secure Poot cancer even to Battlefield V now. If I had any hope that they would consider doing the right thing I would ask for a refund. I will never purchase another piece of digital trash from those rooster sucking creeps ever again. I should sell my EA account and all of the games in it. If someone needs a second EA account and you are willing to put up with the Secure Poot feces being enabled for one of their kids to play independently send me a PM. It's got quite a few titles in it that I will never be installing or playing again.

 

Nice. Forced to enable Secure boot and TPM to play ain't enough anymore. Now you'll need to update the bios as well to be able to play some of the games. What will come next? 

 

If your computer is affected by this bug, you’ll be given a restriction, and you won’t be able to launch Valorant. You can resolve this by updating your motherboard’s BIOS to the latest version and by ensuring that all security features like Secure Boot, VBS, and IOMMU are activated and working properly.

Critical motherboard flaw allows game cheats, Riot Games blocks 'Valorant' players that don't update BIOS

Riot Games, the developer of several popular esports titles such as Valorant and League of Legends, just discovered a security vulnerability affecting several motherboards from Asus, Gigabyte, MSI, and ASRock that cheaters can use to bypass hardware security checks and make it impossible for the game’s anti-cheat software to detect them.

 

 

[Mr. Fox]

5 hours ago, electrosoft said:

It is most likely going to be a UNIFY or APEX.

After owning several MSI boards (Unify X and Carbon) and numerous ASUS boards, if it were me I would stick with ASUS if for no other reason than much better firmware. Apex is a great motherboard generally. Every MSI I have owned has something in the firmware that was irksome to me. They were OK in terms of hardware. Svet deals with fixing lots of MSI firmware messes. They are never-ending. MSI forums are littered with examples of problems caused by screwed up firmware.

4 hours ago, Papusan said:

Nice. Forced to enable Secure boot and TPM to play ain't enough anymore. Now you'll need to update the bios as well to be able to play some of the games. What will come next?

That sucks donkey butt. These idiot game dev clowns are going to FAFO what happens to dictators. They usually end up dead. I am happy that I don't care about Valorant. If I did, I would no longer. It would be dead to me, just as Battlefield and Call of Duty titles that require this stupid nonsense are now. Nasty finger salute to all of them, with both hands. No game is good enough or matters to me to such a degree that I will be told what I have to do in order to play it. They can drop dead and I hope that they do.

[Talon]

4 hours ago, electrosoft said:

Do you plan on using an Apex for some 270k/290k binning?

 

290K only most likely, even if I already know that all Intel has likely done is set D2D and NGU to 32x on both, and slightly binned or overvolted the cores a bit to push clocks. Very likely some sort of Intel APO fluff with new titles as well. If they've somehow fix latency a tad I'd be really shocked but I don't expect it. Still, if I can eek out a couple hundred Mhz on the P/E cores over a 285K, it will be worth it for fun. I imagine we will get a CES launch/announcement with a late Jan street date on the refresh. It sorta blows my mind Intel has waited this long, but it is what it is. 

 

I'm also hoping they launch the Sapphire Rapids HEDT at CES. The cache on those chips looks insane, and honestly it might make for a very expensive gaming CPU if it can be clocked. HEDT was always more fun.