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

[Papusan]

9 hours ago, Mr. Fox said:

GOOD: for only 24 threads, that Cinebench R23 score seems excellent. That is the equivalent of a 5.9-6.0GHz all-P-core and 4.7-4.8GHz all-E-core overclock of a 14900KS. The temperatures and power draw are very good. Will that create overclocking headroom? I don't know. Hopefully it will. It looks like DDR5 overclocking might show improvement. For someone with a Z590 or earlier system, upgrading to this makes more sense than a Z790 equivalent even though it could have been better if they had not ditched hyperthreading. It could have been better than better had they not moved to a stupid tiled chiplet piece of crap and replaced E-core with more hyperthreaded P-cores.

 

Much of this should be put in the box BAD.

 

HT on 8 P-cores can correspond to 6000 points in CBR-23 if you have the thermals and power headroom. Aka a real Arrow lake could pass 50.000 points in CBR-23 with custom cooling and not maxed out OC.

 

10% higher performance with tuned Arrow Lake chips but with same power performance profile as for older 14900KS is less than you should have expected from 3nm vs fake 7nm (actually 10nm, not 7nm) at same power consumption. I mean the CBR-23 performance with no HT aka 24 threads for 285K is bad. Intel's goal this time wasn't better performance. It was all about power efficiency. And that's sad for people with good enough Cpu cooling.

 

Aris is spot on @Mr. Fox

 

 

[Talon]

 

So Intel killed overclocking? You are now limited on voltage with DLVR, cannot exceed a predefined limit in the microcode. If you use DLVR bypass mode, you are limited to 400Mhz unless you're below 10 degrees C. WTF. 

 

Nice job guys, you killed this hobby, OC is done. 

[Mr. Fox]

8 minutes ago, Talon said:

 

So Intel killed overclocking? You are now limited on voltage with DLVR, cannot exceed a predefined limit in the microcode. If you use DLVR bypass mode, you are limited to 400Mhz unless you're below 10 degrees C. WTF. 

 

Nice job guys, you killed this hobby, OC is done. 

If that is how it works out it could be fatal business decision for them. 

[Papusan]

1 hour ago, Talon said:

So Intel killed overclocking? You are now limited on voltage with DLVR, cannot exceed a predefined limit in the microcode. If you use DLVR bypass mode, you are limited to 400Mhz unless you're below 10 degrees C. WTF. 

 

Nice job guys, you kill this hobby, OC is done. 

 

ASUS going this route. Asus going the DELL/Alienware direction... Remove functionality that have been there the whole time. 

 

Remove features in firmware is a bad way to fix problems that not all will experience. See....

2.The option to disable C1E in the BIOS has been removed to ensure it remains enabled, reducing processor stress and maintaining stability per Intel's recommendation.

 

https://www.overclock.net/posts/29379060/ 

[Mr. Fox]

4 hours ago, Papusan said:

 

ASUS going this route. Asus going the DELL/Alienware direction... Remove functionality that have been there the whole time. 

 

Remove features in firmware is a bad way to fix problems that not all will experience. See....

2.The option to disable C1E in the BIOS has been removed to ensure it remains enabled, reducing processor stress and maintaining stability per Intel's recommendation.

 

https://www.overclock.net/posts/29379060/ 

So, was this hard voltage limit and drop to 400 MHz above 10°C another idiotic decision from the dullards at ASSzeus, or a mandate Intel is requiring of all motherboard manufacturers? If this is an ASSzeus decision it is grounds for a boycott, brand-rejection and dismissal from particpation as a computer motherboard manufacturer.

[electrosoft]

13 hours ago, tps3443 said:

I am not sure actually. I’ll check though. 
 

This is my temporary cold bug fix. The cold bug only happens once each day on the very first boot-up from the PC sitting all night/morning. 🤣

 

https://youtu.be/K_JCVLR9YDw?si=bXkyX9OeqrYFO_2B

 

Good to know a variation of ye ole Nintendo Cartridge boot process is still alive and well..... 🤣

 

12 hours ago, Mr. Fox said:

 

I came very close to buying new AMD parts the other day, but this I pulled my head out and remembered how that has turned out every time before and that I have never been satisfied, even once, with anything they make. It's also an engineering abortion with a tiled design and a guaranteed crappy CCD paired with a potentially good or OK CCD. But, the fact that I even paused to momentarily consider doing that shows how disappointing the newer and not-better platform is.

 

 

 

Oh man, this thread would have been filled with some serious fire and brimstone from you if you made that purchase!

 

6 hours ago, Talon said:

 

So Intel killed overclocking? You are now limited on voltage with DLVR, cannot exceed a predefined limit in the microcode. If you use DLVR bypass mode, you are limited to 400Mhz unless you're below 10 degrees C. WTF. 

 

Nice job guys, you killed this hobby, OC is done. 

 

Ugh, this is NOT what we wanted if you enabled DLVR.

 

1 hour ago, Mr. Fox said:

So, was this hard voltage limit and drop to 400 MHz above 10°C another idiotic decision from the dullards at ASSzeus, or a mandate Intel is requiring of all motherboard manufacturers? If this is an ASSzeus decision it is grounds for a boycott, brand-rejection and dismissal from particpation as a computer motherboard manufacturer.

 

Personally, I think this is part of a knee jerk response to the whole 13th/14th response and over compensation to make sure there are no more vmin shifts, mc code problems and other issues. Ugh.....

 

Remember when 13th dropped and the new battle cry was, "memory OC is the new Meta?"....Prophetic....

 

 

[tps3443]

16 minutes ago, electrosoft said:

 

Good to know a variation of ye ole Nintendo Cartridge boot process is still alive and well..... 🤣

 

 

Oh man, this thread would have been filled with some serious fire and brimstone from you if you made that purchase!

 

 

Ugh, this is NOT what we wanted if you enabled DLVR.

 

 

Personally, I think this is part of a knee jerk response to the whole 13th/14th response and over compensation to make sure there are no more vmin shifts, mc code problems and other issues. Ugh.....

 

Remember when 13th dropped and the new battle cry was, "memory OC is the new Meta?"....Prophetic....

 

 

Absolutely. Got to blow some warm air to fire the old beast up on those cold mornings. I may try the other Dominus and see if it has the cold bug as well, or just fix it since I have the parts, but honestly it doesn’t even bother me enough to care, if it did it everytime then yeah, that’d be annoying. It seems to be only when it’s cold and only after slitting all night. I’ve got this thing tuned up great though. 

 

 

[Talon]

6 hours ago, tps3443 said:

Absolutely. Got to blow some warm air to fire the old beast up on those cold mornings. I may try the other Dominus and see if it has the cold bug as well, or just fix it since I have the parts, but honestly it doesn’t even bother me enough to care, if it did it everytime then yeah, that’d be annoying. It seems to be only when it’s cold and only after slitting all night. I’ve got this thing tuned up great though. 

 

 

 

I wonder if you couldn't use some sort of cover or blanket to keep the resistor warmer. It's interesting this was never fixed or found by Asus during validation. I do think this setup and chip were rushed to market though in response to AMD. I remember the memes about it, still watch the video from time to time lol. 

 

 

[Rage Set]

On 10/18/2024 at 10:40 PM, Talon said:

 

I think he will still go hard on them and the platform if it's deserving IMO. 

 

I got the Z890 Apex in today. I also have the OCF that I snagged at $499 on preorder from Newegg before they jacked the price. Will test both I suppose and keep 1. 

 

 

 

Seeing all these leaks though, I'm sorta not feeling good. I'm trying to remain excited and hope it's still at least fun to tune and tinker with, but man, what a let down it seems to be rumored to be. I've said it before, who the F is in charge over there? Why do we have a useless NPU on a desktop CPU that will be paired with a huge RTX 4090 or 5090? Why do I need an NPU with a useless 13 tops? I mean the 5800X3D came out like 3 years ago? Why the F didn't Intel say dang, lets up the cache and slap it on hard. I know cache has it's drawbacks and penalties, but it seems to really help with latency prone chiplets/tiles. 

 

Nope, Intel said give them a totally useless NPU, that will make all the reviewers blush.

 

 

Outside of the idiotic Intel protections, my doubts about the 285K and the 1851 as a whole have tempered a bit. I think Intel allowing the E cores to potentially overclock like crazy is the saving grace. 

 

I am very intrigued about this platform now. 

 

EDIT: Guys, I think we need to relax. OC'ing isn't dead, at least not for guys like us. We will still be able to extract more performance out of these chips and some of the "guardrails" will be removed when Intel realizes people aren't happy.

 

In Jay's video, you noticed he said DLVR Bypass isn't available at launch. That means Intel is likely taking a wait and see approach.

 

Give it some time and I say, wait for the reviews. 

[Papusan]

 

G. Skill just hate spend money on thermal pads. Also the Clock driver don't get any thermal pad. How much did G. Skill save for an 3cm² thermal pad ? 10-20 cent? This is real greed.

 

 

 

Asus, what the hell is going on with the pricing.... The ROG tax ain't enough anymore. 

 

Quote

Being the Guinea pig is damn expensive. But this is normal with AMD products. First maximize the profits, then reduce it afterwards.

 

AMD make ready for the Arrow Lake launch. I'm quite sure AMD pay the price difference between the MSRP and the lower retail prices.  

AMD Ryzen 9 9950X drops to $599 for the first time, 9900X now 14% lower at $429

 

 

[tps3443]

5 hours ago, Talon said:

 

I wonder if you couldn't use some sort of cover or blanket to keep the resistor warmer. It's interesting this was never fixed or found by Asus during validation. I do think this setup and chip were rushed to market though in response to AMD. I remember the memes about it, still watch the video from time to time lol. 

 

 

Just typical Asus and their poor QA. Two resistors were literally soldered on backwards lol. Not the first batch, but later batches had this issue apparently.
 

When the boards were sent in for repair they were showing up in warmer environments for repair. And the problems were completely hidden inside repair facilities (“Asus repair would say “nothing wrong with motherboard send back”) I can only assume since the ambient temps were probably warmer. Consumers could not fire their systems in their cooler homes, then the customer would think it must be my CPU that is dead, let me spend another $3,000 and send this W3175X back 🤯This little mistake caused many to think their boards or CPU’s were broken with Code 00 showing CPU as the problem lol. One quick blast with some warm air will fire it up. Instead people were tearing down their rigs and sending parts for exchange/repair. It was a bad experience I can only imagine. Wendell from Level1Tech who discovered this issue is still running his W3175X system. The person to blame is definitely Asus. Because the Dominus was the most popular choice. The Aorus C621 Extreme has less than #50 motherboards made, they made some more with a total of less than #100 retail boards in the wild, only they were without any issues with this to report. SR3 Dark came later, no issues either as far as I know. 
 

This cold bug was also present on some Asus Rampage Extreme X99/X299 boards and other boards too. ( Asus Quality Assurance at its finest) the good news is though, I don’t think this platform has any actual issues other than the cold bug. Any others problems are gonna be user error, throwing in random ram, and also improper cpu installation. People were not reading the manual and guessing on which 8 pin/24 pin connectors to plug in. 
 

As for my setup, I can only go by what I experience. Not a single crash, no problems other than the cold bug I already knew about. 🤗 I believe it to be more stable than LGA1700 so far. 
 

PS: No cold bug today. I just turned my rig on. 😎

[Mr. Fox]

8 hours ago, Rage Set said:

 

Outside of the idiotic Intel protections, my doubts about the 285K and the 1851 as a whole have tempered a bit. I think Intel allowing the E cores to potentially overclock like crazy is the saving grace. 

 

I am very intrigued about this platform now. 

 

EDIT: Guys, I think we need to relax. OC'ing isn't dead, at least not for guys like us. We will still be able to extract more performance out of these chips and some of the "guardrails" will be removed when Intel realizes people aren't happy.

 

In Jay's video, you noticed he said DLVR Bypass isn't available at launch. That means Intel is likely taking a wait and see approach.

 

Give it some time and I say, wait for the reviews. 

It is very hard not to be intrigued by it if Cinebench R23 multi-core results are truly matching an overclocked 13900/14900/KF/KS CPU in spite of the missing hyperthreading. I find that hard to believe and I am skeptical, but intrigued if it turns out to be true. At the same time it bugs the crap out of me wondering how much better it would have been had they not made so many idiotic decisions. The mere mention of the NPU part makes me angry. I am 100% anti-AI and I don't like having cockamamie Orwellian sci-fi crap like that thrust upon us.

 

The ASRock Taichi OC Formula is also an intriguing motherboard. I almost pre-ordered that motherboard to have it sitting on the shelf just in case, but that would be silly because I am still taking a wait-and-see approach here on the CPU part. Everything else looks amazing about Z890 except for the low CPU core/thread count and disgusting tile design. This will definitely be a platform to be extremely cautious about applying firmware updates so you don't end up with an emasculated door stop that is worthless as tits on a bull if you are an overclocker.

7 hours ago, Papusan said:

 

G. Skill just hate spend money on thermal pads. Also the Clock driver don't get any thermal pad. How much did G. Skill save? 10 cent ?  This is real greed.

 

 

 

Asus, what the hell is going on with the pricing.... The ROG tax ain't enough anymore. 

 

 

AMD make ready for the Arrow Lake launch. I'm quite sure AMD pay the price difference between the MSRP and the lower retail prices.  

AMD Ryzen 9 9950X drops to $599 for the first time, 9900X now 14% lower at $429

 

 

I do not think G.SKILL has ever paid attention to memory thermals. They have always been the worst when it comes to deliverying cruddy thermal solutions on their memory modules. They are similar to ASSzeus in that they target overclocking enthusiast and then cut corners in some areas that matter. They're both too hung up on aesthetics and RGB rainbow puke, and even there their design teams have kind of tacky aesthetic preferences. As far as ASSzeus goes, their approach to things has never been more questionable than it is today. It's no secret that I haven't liked the brand for a long time, but it's getting easier than ever to hate them. Most of it is self-inflicted on their part. Bad people + bad decisions = bad outcomes. The bad pricing of their junk adds insult to injury.

15 hours ago, electrosoft said:

Oh man, this thread would have been filled with some serious fire and brimstone from you if you made that purchase!

Being a PC hardware junky is tough because you want to learn from your mistakes and never repeat them again, but you always find things intriguing no matter what the brand is. You always want something awesome regardless of who made it. The problem is easily exacerbated by deceptively clever marketing (often blatant lies and deliberate misrepresentations) and fanboys that never speak negatively about products made by their preferred brand. For whatever reason, it seems like many (not all) AMD fanboys are more inclined to show unconditional love for AMD while Intel fanboys typically hold back no hate when expressing contempt for Intel's mistakes.

 

I am a forgiving person to a fault and forgiving people is essential. Harboring resentment against individuals is wrong and harmful to all involved; most of all to the person holding resentment and unforgiveness. I have to forcefully remind myself that forgiving companies, politicians and political parties is extremely foolish. Remembering everything they have done wrong, and reminding others of their shortcomings and failures is wise and essential to not getting screwed by them again in the future. Knowing that skunks don't change their stripes, obtaining forgiveness and finding a path to redemption for them (as corporate bodies) should be extremely difficult, if not nearly impossible.

[tps3443]

I’m finally resetting Windows. There are about 300 running services too many for me on my current OS lol. This is an old Windows maybe like 1 year old now. I have used it for work, benching, gaming, modding, creating, and breaking, and fixing things all in one OS. Not to mention the CPU/RAM/platform swaps. 
 

Nothing quite like a fresh OS. You get all that storage space back that you could never find. 🤣

[Mr. Fox]

18 minutes ago, tps3443 said:

I’m finally resetting Windows. There are about 300 running services too many for me on my current OS lol. This is an old Windows maybe like 1 year old now. I have used it for work, benching, gaming, modding, creating, and breaking, and fixing things all in one OS. Not to mention the CPU/RAM/platform swaps. 
 

Nothing quite like a fresh OS. You get all that storage space back that you could never find. 🤣

You should try this. I have found it to perform a lot better than GhostOS. Far less bloat on this one. I obviously prefer Windoze 10 (22H2) over Winduhz 11. I would recommend avoiding versions past 22H2 for either OS.

 

https://windowsxlite.com/

[tps3443]

47 minutes ago, Mr. Fox said:

You should try this. I have found it to perform a lot better than GhostOS. Far less bloat on this one. I obviously prefer Windoze 10 (22H2) over Winduhz 11. I would recommend avoiding versions past 22H2 for either OS.

 

https://windowsxlite.com/

I use a regular Windows 11 Pro “Full retail install” I had to buy it about 6 months ago. I love the lite Windows that are available, but I stopped using them a while back. I noticed memory stability issues with some of them, and also certain work apps that I need won’t work for me. Instead of trouble shooting each app, I just went back to a regular windows 11 Pro install and I will disable or remove most of the stuff I don’t need or use.

[Mr. Fox]

2 hours ago, tps3443 said:

I use a regular Windows 11 Pro “Full retail install” I had to buy it about 6 months ago. I love the lite Windows that are available, but I stopped using them a while back. I noticed memory stability issues with some of them, and also certain work apps that I need won’t work for me. Instead of trouble shooting each app, I just went back to a regular windows 11 Pro install and I will disable or remove most of the stuff I don’t need or use.

You can also multi-boot. I have a fully metastasized cancer version of Winduhz 11 that I use when I need to or want to test something sketchy or bloaty, then I have optimized Window(s) installations that prioritize performance above all else. I have to use a cancer version for work because I use PowerBI daily and my optimized OSes are missing the feces that PowerBI needs to function. The 24H2 version is about as cancerous, bloated and rancid as it can get.

[Clamibot]

3 hours ago, Mr. Fox said:

You should try this. I have found it to perform a lot better than GhostOS. Far less bloat on this one. I obviously prefer Windoze 10 (22H2) over Winduhz 11. I would recommend avoiding versions past 22H2 for either OS.

 

https://windowsxlite.com/

 

I'm definitely gonna try that out as I'm starting to run into sofware compatibility issues (both for games and work) with my V1809 LTSC install, so I now have a genuine need for a newer version of windows. Thanks for posting this!

 

I also got an Asus XG309CM monitor and absolutely love the 220 Hz refresh rate. I don't like that the max refresh rate isn't perfectly divisible by 24 or 30 though so I've been looking into some monitor overclocking (which I've done before), but am running into a bit of a snag this time. Apparently, there's a refresh rate limiter on this monitor according to CRU. Does anyone know how to bypass such a thing? I've never seen something like this before on a monitor. I just want to overclock it to 240 Hz, which I think will be doable on this panel.

[tps3443]

The Wifi on this Dominus is beastly! Not sure why, but I am hitting 1Gbps in steam downloading games which is really great. But I do have some issues with the Bluetooth working on the Dominus Extreme. I have re-installed Windows, tried multiple different BT/Wifi drivers, gone through device manager adding new drivers, uninstalling/re-installing etc, and I cannot get BT working for the life of me. It's enabled in the bios, and I can see it pick up a device like my Xbox controller in Windows 11, but I cannot actually pair the Xboxcontroller. It just says something like "Device failed to connect" Been at this for about a day now lol. Any ideas? Anyone ever had this happen to them? 

I am thinking of just pulling the Wifi/BT 9260 off the board all together, and putting a different one in. I feel like I have tried it all. 


EDIT: This is a nice 4.6Ghz run on all cores. Still very good CPU temps. Power is 528 watts MAX. I really love this 3175X its really a sweet chip. 😍 

[Mr. Fox]

9 hours ago, Clamibot said:

 

I'm definitely gonna try that out as I'm starting to run into sofware compatibility issues (both for games and work) with my V1809 LTSC install, so I now have a genuine need for a newer version of windows. Thanks for posting this!

 

I also got an Asus XG309CM monitor and absolutely love the 220 Hz refresh rate. I don't like that the max refresh rate isn't perfectly divisible by 24 or 30 though so I've been looking into some monitor overclocking (which I've done before), but am running into a bit of a snag this time. Apparently, there's a refresh rate limiter on this monitor according to CRU. Does anyone know how to bypass such a thing? I've never seen something like this before on a monitor. I just want to overclock it to 240 Hz, which I think will be doable on this panel.

You're welcome.

 

I've never encountered that refresh rate limited thing before. At least not that I recognized. I will check my ASUS 4K 120Hz monitor and see if CRU shows the same thing.

 

My cheap Philips 1080p 100Hz monitor is running fine at 2560x1440 100Hz. I use this monitor primarily for having HWiNFO64 Sensor window open when benching or gaming.

[Papusan]

On 10/20/2024 at 3:57 PM, Rage Set said:

I think Intel allowing the E cores to potentially overclock like crazy is the saving grace. 

 

Yup. Without real overclock on the baby-cores this looks worse and worse if the results is real. Core Ultra 9 285K already use +370W with unlimited power and barely manage 46.000 in CBR-23. That is on level with an nice binned oc'd 14th gen Core i9. 

 

This looks darn disgusting... CBR23 ~ 46289@370W. And the ring is heavly castrated. Measly 3.8GHz. I wonder how far you can go with a Ring clock speed running at Jokebook clocks.

 

 

https://videocardz.com/pixel/unlocked-intel-core-ultra-9-285k-approaches-370w-power-draw-during-cinebench-test

[Mr. Fox]

11 minutes ago, Papusan said:

 

Yup. Without real overclock on the baby-cores this looks worse and worse if the results is real. Core Ultra 9 285K already use +370W with unlimited power and barely manage 46.000 in CBR-23. That is on level with an nice binned 14th gen Core i9. 

 

This looks darn disgusting... CBR23 ~ 46289@370W. And the ring is heavly castrated. Measly 3.8GHz. I wonder how far you can go with a Ring clock speed running at Jokebook clocks.

 

 

https://videocardz.com/pixel/unlocked-intel-core-ultra-9-285k-approaches-370w-power-draw-during-cinebench-test

The new Core Ultra 9 285K could have been better in a lot of ways, and it is pretty clear we agree on that. I am sure there are going to be caveats and strings attached that we don't know about yet.

 

That being said, how is having a higher score with 8 fewer threads to handle the workload a bad thing? Setting aside how much better it could have been had they not eliminated hyperthreading, this is a substantially higher Cinebench score compared to a 14900K/KF/KS with a significant overclock at much higher voltage, power draw and heat production. It looks like it is probably going to beat Gen 13 and 14 CPUs and the latest and greatest Ryzen flagship. While we still have things to complain about, is that not a win on face value? Hitting 46K with a nicely binned 14900KF/KS requires a severe overclocks and abnormally excellent cooling with sub-ambient temperatures.

[tps3443]

1 hour ago, Mr. Fox said:

The new Core Ultra 9 285K could have been better in a lot of ways, and it is pretty clear we agree on that. I am sure there are going to be caveats and strings attached that we don't know about yet.

 

That being said, how is having a higher score with 8 fewer threads to handle the workload a bad thing? Setting aside how much better it could have been had they not eliminated hyperthreading, this is a substantially higher Cinebench score compared to a 14900K/KF/KS with a significant overclock at much higher voltage, power draw and heat production. It looks like it is probably going to beat Gen 13 and 14 CPUs and the latest and greatest Ryzen flagship. While we still have things to complain about, is that not a win on face value? Hitting 46K with a nicely binned 14900KF/KS requires a severe overclocks and abnormally excellent cooling with sub-ambient temperatures.

One area where these chips are weaker is, doing things on the desktop like work apps running, and then going to play some games at the same time. The 285K chip would really need everything closed out on the desktop for best absolute performance in games, maybe even a quick restart. There are only 8 P-Cores. Anything moderately intensive running on the desktop or in the background will tie these cores up for sure. I know this all too well, because I run a gazillion apps at once, and go play games simultaneously while I am waiting on one of my agents to join a meeting or I'm just killing time in general. The 285K only has "24 threads" so definitely.  All of this makes me really miss the old days of hardware. MOAR COREZ era lol. 

[Mr. Fox]

1 hour ago, tps3443 said:

One area where these chips are weaker is, doing things on the desktop like work apps running, and then going to play some games at the same time. The 285K chip would really need everything closed out on the desktop for best absolute performance in games, maybe even a quick restart. There are only 8 P-Cores. Anything moderately intensive running on the desktop or in the background will tie these cores up for sure. I know this all too well, because I run a gazillion apps at once, and go play games simultaneously while I am waiting on one of my agents to join a meeting or I'm just killing time in general. The 285K only has "24 threads" so definitely.  All of this makes me really miss the old days of hardware. MOAR COREZ era lol. 

I think any time something is marketed as a "gaming" [insert item here] it is secret code for probably sucky at everything else.

 

I've never played games while simultaneously doing something else before, but I know not everyone does things the same way. I wouldn't even try to do that because I would feel like I was losing my marbles if I did. I can't walk at chew gum at the same time, but I excel at doing one or the other.

[Talon]

 

Go time boys. 

[tps3443]

33 minutes ago, Talon said:

 

Go time boys. 

That motherboard looks so familiar for some reason 🤣