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

[electrosoft]

17 minutes ago, Talon said:

 

Had they just gone with 2x 12v2x6 connectors, it would have been worth it for the niche market IMO. They would have instantly sold out. Nobody really wants to have to buy a BTF board, at least on Intel side and have worse memory performance over the 2x DIMM OC boards. Maybe next gen will have BTF on 2x DIMM boards, but that is a gamble and who really wants to wait. 

 

This entire gen has been a mess. I got my waterblock for my HOF XOC card from Ali delivered this week. I still have not gotten the rest of the parts but plan to have those soon and can finally put this bad boy under water. In the end, I think I'll still perfer a good air cooler. But given the recent news of 2027 for RTX 6000 series, I will at least get to get some use out of the setup before it's obsolete lol. 

 

Agreed, AI has totally shifted priorities this gen along with Nvidia exerting more and more control each generation along with pricing is getting painful. Jensen little man syndrome is in full effect unfortunately.

 

I would hate to pick up that Asus 5090 (as beautiful and well crafted as it is) and be locked into certain motherboards going forward. I can't imagine not being able to test it in other systems easily which is also one reason I moved away from AIO based GPUs to air cooled this generation having previously used the 3090 Kingpin, 3090ti Kingpin and then Suprim 4090 Liquid. Swapping into open frame MBs was a nightmare with that AIO.

 

I saw where they are now saying Super cards aren't coming till Q3 2026 which does not bode well for the 6000 series launch. If that is true, then yeah 6000 series won't launch until well into 2027.

 

If rumors of greatly reducing or ceasing production and shifting the top end to the 5080 comes true that is even a darker portend of things to come for us mere consumers....

 

 

[Mr. Fox]

53 minutes ago, electrosoft said:

Agreed, AI has totally shifted priorities this gen along with Nvidia exerting more and more control each generation along with pricing is getting painful. Jensen little man syndrome is in full effect unfortunately.

 

I would hate to pick up that Asus 5090 (as beautiful and well crafted as it is) and be locked into certain motherboards going forward. I can't imagine not being able to test it in other systems easily which is also one reason I moved away from AIO based GPUs to air cooled this generation having previously used the 3090 Kingpin, 3090ti Kingpin and then Suprim 4090 Liquid. Swapping into open frame MBs was a nightmare with that AIO.

 

I saw where they are now saying Super cards aren't coming till Q3 2026 which does not bode well for the 6000 series launch. If that is true, then yeah 6000 series won't launch until well into 2027.

 

If rumors of greatly reducing or ceasing production and shifting the top end to the 5080 comes true that is even a darker portend of things to come for us mere consumers....

A$$zeus selling GPUs that can only work with one of their motherboards was likely a very deliberate act and solely the result of ulterior motives. Not a smart buy for anyone that hasn't sold their soul to the ROG clown posse. The idea itself isn't a bad one. Much better than a fragile arson 12VHPWR (aka 12V-2x6) cable. The proprietary part makes it suck.

 

With the baloney we are seeing with respect to AMD/Radeon GPUs now, it just gives NVIDIA a stronger chokehold on the GPU market that ultimately benefits only NVIDIA. There is no reason for them to feel compelled to release a Super or 6000 GPU line because they are effectively only competing with themselves. You either take what they offer and pay more than it is worth, or you settle for something substantially less. If AMD doesn't keep their prices in check they'll have nothing to sell to budget-conscious shoppers that are willing to sacrifice performance and good thermals to save money.

[Papusan]

3 hours ago, Talon said:

Had they just gone with 2x 12v2x6 connectors, it would have been worth it for the niche market IMO. They would have instantly sold out. Nobody really wants to have to buy a BTF board, at least on Intel side and have worse memory performance over the 2x DIMM OC boards. Maybe next gen will have BTF on 2x DIMM boards, but that is a gamble and who really wants to wait. 

 

You'll need newer pc chassis as well to fit the BTF motherboard. But You may fix that with a dremel and some paint but Asus main point is that you should buy whole new hardware only for that extra power connector. 

 

Greedy.......

[jaybee83]

4 hours ago, electrosoft said:

Exactly. With the WV2P, I will move away from the MSI tentacle. For my own edification, I do plan on running it for at least a week so I can monitor what's going on.

 

How are 5090 stock levels there? They are drying up so quickly over here and back up to high pricing in rapid fashion.

 

so far no too shabby tbh with regards to 5090 availability and pricing. total of 41 models on offer, out of which 33 are in stock and ready for shipping. another 3 skus need a bit of a wait time and only 5 are currently out of stock.

 

currently cheapest model, the Inno3D X3 goes for 2399€, thus far only a "medium" increase over its all time low price point at 2076€.

 

the Suprim SOC had its lowest price back in Sep until mid-Oct at 2640€, has since slightly increased to 2840€. The Astrai had a low at 2740€ back in Aug and has now increased to 2880€, so its been more stable than other models, price curve still pretty flat but at a much higher price level than other models.

 

the upcreep is slight thus far but one can discern a slow upward trend. still time to snatch a model at a comparably decent price tho.

 

from what i can discern, the cheapest ever available 5090 model was the Ventus back in June at 1999€.

[Papusan]

 

Still more than enough 5090's available here home. Cheapest one is GIGABYTE GeForce RTX 5090 WindForce 3 OC at $1988,15 MSRP (without tax). So right below the famous 5090FE MSRP. 

 

 

New rumors. 5000 series Super cards just delayed. Now planned to launch in 26Q3

NVIDIA GeForce RTX 50 SUPER reportedly slips to Q3 2026, RTX 5060 Ti 16GB in short supply soon

[Mr. Fox]

1 hour ago, Papusan said:

 

Still more than enough 5090's available here home. Cheapest one is GIGABYTE GeForce RTX 5090 WindForce 3 OC at $1988,15 MSRP (without tax). So right below the famous 5090FE MSRP. 

 

 

New rumors. 5000 series Super cards just delayed. Now planned to launch in 26Q3

NVIDIA GeForce RTX 50 SUPER reportedly slips to Q3 2026, RTX 5060 Ti 16GB in short supply soon

That is a good price. I would replace my 4090 with a Windforce 5090 if I could buy it that price here. I'd like to try the mod adding the second 12VHPWR connection to the empty space on the PCB. The cheapest I have noticed here is $1999 for the Zotoc 5090 Solid (non-OC).

 

Edit: Hmm. Looks like it is. 

 

https://www.microcenter.com/product/692139/gigabyte-nvidia-geforce-rtx-5090-windforce-triple-fan-32gb-gddr7-pcie-50-graphics-card

 

Edited November 8 by Mr. Fox
windfarts 5090

[Mr. Fox]

Another good and valid reason to build you own and not buy prebuilt garbage. It would be interesting to see if forcing a normal ASUS BIOS using an SPI flash would work.

 

 

[Mr. Fox]

Happy Saturday. Here's some humor to brighten your weekend.

[Mr. Fox]

20 hours ago, Papusan said:

 

You'll need newer pc chassis as well to fit the BTF motherboard. But You may fix that with a dremel and some paint but Asus main point is that you should buy whole new hardware only for that extra power connector. 

 

Greedy.......

The problem is way too much of the new tech exists for all of the wrong reasons. Optics, aesthetics and form over function. Not better, stronger, faster, more reliable. Change purely for the sake of change, not because it was needed. They need something shiny and new to sell, and otherwise have nothing to offer. It doesn't actually need to be any good if the sheeple believe it is and buy it.

[Mr. Fox]

I finally took the time to get the X870E-E Strix memory on water like the Apex. First time in quite a long time that I ran the memory with air cooling and I am so glad to not have to use a fan for it any more. I really did not like that. Water is so much better, not to mention looking a whole lot better as well. That 120mm fan hanging on a bracket from the top radiator was pretty ugly.

 

It will take a day or two to work all of the air out of the distro block.

 

Now it's time to hit the sack.

 

[Meaker]

6 hours ago, Mr. Fox said:

I finally took the time to get the X870E-E Strix memory on water like the Apex. First time in quite a long time that I ran the memory with air cooling and I am so glad to not have to use a fan for it any more. I really did not like that. Water is so much better, not to mention looking a whole better as well. That 120mm fan hanging on a bracket from the top radiator was pretty ugly.

 

It will take a day or two to work all of the air out of the distro block.

 

Now it's time to hit the sack.

 

Nicely an clean even with the extra pipes, the distribution plate puts in some good work.

I got my network into it's final form. Full 10Gb backbone to my study and living room with 2.5Gb connections.

 

DAC cable to the 8 port switch and an SFP+ to ethernet adaptor for 10Gb for my uplink to upstairs.

I'm getting a revised version of this holder so I can cable tidy this properly.

[Mr. Fox]

6 hours ago, Meaker said:

Nicely an clean even with the extra pipes, the distribution plate puts in some good work.

I got my network into it's final form. Full 10Gb backbone to my study and living room with 2.5Gb connections.

 

DAC cable to the 8 port switch and an SFP+ to ethernet adaptor for 10Gb for my uplink to upstairs.

I'm getting a revised version of this holder so I can cable tidy this properly.

Thank you. Adding memory into a cooling loop is much aesthetically cleaner with a distribution block or manifold. I do not care for it when trying to tap in from the CPU or GPU block. It looks too cluttered and busy for my taste. The connections look all jammed together and chaotic to me.

 

You'e got the network configured nicely. I bet you can feel a big improvement with how things work as well. Our home network is not nearly as sophisticated or organized, but when we had the Gigabit fibre installed by Quantum they included a couple of network mesh "pods" at no cost and having those in place dramatically improved wireless connectivity for devices at the opposite end of the home. Our home is not that big (1800 square feet, single story) but it still helped a lot.

 

 I do not know what is going on with imgbb hosting but I can hardly use their service at this point. Photos take forever (or sometimes never) to load and often show broken links. It's not user error, but something with their servers. Even viewing uploaded images from within my account on their web site it is the same.

[Meaker]

21 minutes ago, Mr. Fox said:

Thank you. Adding memory into a cooling loop is my aesthetically cleaner with a distribution block or manifold. I do not care for it when trying to tap in from the CPU or GPU block. It looks to cluttered and busy for my taste. The connections look all jammed together and chaotic to me.

 

You'e got the network configured nicely. I bet you can feel a big improvement with how things work as well. Our home network is not nearly as sophisticated or organized, but when we had the Gigabit fibre installed by Quantum they included a couple of network mesh "pods" at no cost and having those in place dramatically improved wireless connectivity for devices at the opposite end of the home. Our home is not that big (1800 square feet, single story) but it still helped a lot.

 

 I do not know what is going on with imgbb hosting but I can hardly use their service at this point. Photos take forever (or sometimes never) to load and often show broken links. It's not user error, but something with their servers. Even viewing uploaded images from within my account on their web site it is the same.

The big thing is when my partner is doing anything with the NAS it has no impact on my own experience now, any one person can max out their connection and the rest of the network is just unaffected.

[Papusan]

 

Be aware of possible ticking bomb as decreased performance. Being the guinea pig is never a good idea.

Intel stomps down 30 bugs including privilege escalation vulnerabilities

Intel releases new microcode with "hidden" security patches to fix 30 vulnerabilities in nearly 200 CPUs: issues with iGPU, Thread Director, Ethernet, WiFi, and more

[Mr. Fox]

Nice new feature in ASUS BIOS for F3 Save as CMOS file to save all BIOS files to USB as a single file instead of a separate CMO file for current profile only.

 

Explanation from safedisk: https://www.overclock.net/posts/29527571/

Quote

safedisk said: 

*New features added
CMO will save current BIOS setting as .CMO file
CMOS will save all exist profiles to single .CMOS file

So if you save it as a cmos file, you can import all presets with one file

@jaybee83 @Raiderman https://www.overclock.net/posts/29527497/

 

Explains why CMOS is much larger than CMO.

[jaybee83]

1 hour ago, Mr. Fox said:

Nice new feature in ASUS BIOS for F3 Save as CMOS file to save all BIOS files to USB as a single file instead of a separate CMO file for current profile only.

 

Explanation from safedisk: https://www.overclock.net/posts/29527571/

@jaybee83 @Raiderman https://www.overclock.net/posts/29527497/

 

Explains why CMOS is much larger than CMO.

 

nifty, but my main gripe is still that in many instances i need to manually come up with new profiles whenever i update the bios version. cmon now, how hard can it be to make bios profiles compatible with all new bios versions?!

[Mr. Fox]

17 minutes ago, jaybee83 said:

 

nifty, but my main gripe is still that in many instances i need to manually come up with new profiles whenever i update the bios version. cmon now, how hard can it be to make bios profiles compatible with all new bios versions?!

I have been able to use my CMO file from the prior BIOS version, but the trick is to save a profile with the new BIOS first, then apply the old profile over it. (Safedisk shared this.) So after flashing I disable TPM, Secure Poot, iGPU, WiFi, BT and save it as a new BIOS profile. F10 to save and exit, then go back in and apply my OC CMO profile from the previous BIOS. That has not given me any issues. After applying the old profile and confirming all was well, then save again as a new OC profile and new CMO file for the updated BIOS version.

 

I could see where an old profile might be incompatible if they removed, changed or added new features or rearranged menu order, but if the only changes are underlying code and BIOS default values there is no excuse for them not being compatible.

 

I ordered an EPYC 4585PX. I hope it is better at core and memory overclocking than either of my average 9950X so I don't have to RMA/refund yet another one. The only reason I did not return either of the two I have now is they were average based on what I could tell looking at other samples judging from HWBOT scores and better than the below average trash samples I RMA'd before them.

 

If it is not better than I am probably just going to just send it back for refund and be done with trying to enjoy lackluster AM5 overclocking. The 9950X3D that I returned was the worst Ryzen 9 sample I have ever seen. It was an absolute POS.

 

Even gaming it turning to crap. Having to enable TPM and Secure Poot filth to play the new BF and CoD releases is totally unacceptable. I about popped a vein when I discovered they had retroactively applied Javelin to BF 2042 and rendered a game I had thoroughly enjoyed totally worthless to me. Bastards.

[electrosoft]

6 hours ago, jaybee83 said:

 

nifty, but my main gripe is still that in many instances i need to manually come up with new profiles whenever i update the bios version. cmon now, how hard can it be to make bios profiles compatible with all new bios versions?!

 

Same, a function to import old BIOS settings would be nice. I've just gotten used to screen shotting my settings at this point and then manually re-entering them. Doesn't take too long (~10min tops if that) but still a nuisance. 

 

6 hours ago, Mr. Fox said:

I have been able to use my CMO file from the prior BIOS version, but the trick is to save a profile with the new BIOS first, then apply the old profile over it. (Safedisk shared this.) So after flashing I disable TPM, Secure Poot, iGPU, WiFi, BT and save it as a new BIOS profile. F10 to save and exit, then go back in and apply my OC CMO profile from the previous BIOS. That has not given me any issues. After applying the old profile and confirming all was well, then save again as a new OC profile and new CMO file for the updated BIOS version.

 

Hmmmm, I'll have to give this a true next time on the z690 D4 strix w/ SP109 14900KS and x870e Hero....

 

6 hours ago, Mr. Fox said:

I ordered an EPYC 4585PX. I hope it is better at core and memory overclocking than either of my average 9950X so I don't have to RMA/refund yet another one. The only reason I did not return either of the two I have now is they were average based on what I could tell looking at other samples judging from HWBOT scores and better than the below average trash samples I RMA'd before them.

 

If it is not better than I am probably just going to just send it back for refund and be done with trying to enjoy lackluster AM5 overclocking. The 9950X3D that I returned was the worst Ryzen 9 sample I have ever seen. It was an absolute POS.

 

 

Didn't even know they basically offered the 9950X3D in Epyc form. Looks like the few reviews and look ups say it is better binned than the 9950X3D and the price is now down to basically 9950X3D levels. Tempted to try one myself! 

 

6 hours ago, Mr. Fox said:

Even gaming it turning to crap. Having to enable TPM and Secure Poot filth to play the new BF and CoD releases is totally unacceptable. I about popped a vein when I discovered they had retroactively applied Javelin to BF 2042 and rendered a game I had thoroughly enjoyed totally worthless to me. Bastards.

 

Unfortunately cheaters and cheat suppliers have gotten so good that EA has implemented their version of "The Final Solution" and it works. It wipes out any and all chances of cheats sitting in any memory space without being detected but it requires you to basically give up just about every aspect of protection on your system to do it. It is insane watching cheaters lose their minds in threads because they are being banned left and right trying to run updated paid for (!) cheats that have always worked getting them permanently banned on a hardware level. Just like with the Switch2, buyers have to be wary buying second hand hardware that could be banned from the gaming networks or games they want to play.

 

I am not a fan of the level of access that is required to play some of the newest competitive games, but I also understand why in such a hyper competitive environment. When I used to play competitive Quake the sheer amount of cheaters was ridiculous and was always refreshing at gaming cons to suddenly see many of those same players under observation during matches suddenly get "less good" real quick. 

 

Only way I would entertain this level of intrusion to play would be a separate, games only install of Windows with zero Windows login or any other type of logins anywhere and use it to game exclusively on its own drive and I'm not willing to go to that level yet.

 

Next WoW xpac, Blizzard is basically taking security to the next level and locking out mods and assists in what appears to be an attempt to heighten security. Remember, Blizzard via Activision was purchased for almost 70billion from Micro$oft. The Midnight xpac is supposed to be another major overhaul so we will see.

 

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

 

I like Testing Games to a degree, but the fact they still test the 14900k unoptimized at 6000 memory vs the X3D running the classic sweet spot 6000 just doesn't work for me. The way they run their DDR5 my tuned DDR4 B-die setup with the 14900KS will decimate their results each and every time let alone when I was running it tuned w/ 8400 DDR5.....

 

 

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

 

In the same vein of Socket 1700 and BF, if the 12 core Bartlett-S drops and is tunable, I'm still definitely interested in it too depending on where we can go with it. I'd pick up an extra board and slap these 8400 kingspecs in there and  see where it goes....

 

 

[Raiderman]

11 hours ago, jaybee83 said:

 

nifty, but my main gripe is still that in many instances i need to manually come up with new profiles whenever i update the bios version. cmon now, how hard can it be to make bios profiles compatible with all new bios versions?!

Agreed! 

[Talon]

20 hours ago, electrosoft said:

I like Testing Games to a degree, but the fact they still test the 14900k unoptimized at 6000 memory vs the X3D running the classic sweet spot 6000 just doesn't work for me. The way they run their DDR5 my tuned DDR4 B-die setup with the 14900KS will decimate their results each and every time let alone when I was running it tuned w/ 8400 DDR5.....

 

They fact that they used basically the best memory option 6000 CL30 for the AMD part, but then use far below Intel's optimal spec is enough for me to call it biased. 

 

Even if you use a mainstream 4 dimmer, most can achieve around 7600 CL32-34. Hell even using 7000 CL30 2x32gb dual rank would be a better option as it will offer up gaming performance of an 8000 SR kit on Intel. 

 

Unfortunately, this type of testing is mainstream and how most outlets test. It is what it is, I've just learned to ignore most of it. 

[electrosoft]

2 hours ago, Talon said:

 

They fact that they used basically the best memory option 6000 CL30 for the AMD part, but then use far below Intel's optimal spec is enough for me to call it biased. 

 

Even if you use a mainstream 4 dimmer, most can achieve around 7600 CL32-34. Hell even using 7000 CL30 2x32gb dual rank would be a better option as it will offer up gaming performance of an 8000 SR kit on Intel. 

 

Unfortunately, this type of testing is mainstream and how most outlets test. It is what it is, I've just learned to ignore most of it. 

 

Exactly. I always reserve judgement till I see the testing parameters then adjust accordingly. I think the 285k gets a bad wrap too:

 

Frame Chasers and Blackbird Tech are two of the few who tune their CPUs and memory for testing to at least show Intel in the proper light for those who truly care where the chips fall when optimized.

 

 

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

 

FE / 6000 connector fragility strikes again and the problem is unable to source the part.....

 

 

[Papusan]

 

Nice. The new Intel flagship got one boost bin upgrade of 100 MHz for TVB. But no change in all core boost clocks(Turbo boost max). Everything else is almost exactly the same. What a fantastic upgrade. Run and buy. Why bother put resources on an refresh? Plus models? LOL

 

 

Intel Core Ultra 290K, 270K and 250K Plus spec leak: “Arrow Lake Refresh” with higher clocks, more cores and faster memory support

 

Core Ultra 200K Plus (overview).

Core Ultra 9 285K➡️ 290K Plus: + 100 MHz P-Core (Max)

Core Ultra 7 265K➡️ 270K Plus: + 4 E-Cores

Core Ultra 5 245K➡️ 250K Plus: + 100 MHz P-Core (Max), + 4 E-Cores

[Mr. Fox]

14 minutes ago, Papusan said:

Nice. The new Intel flagship got one boost bin upgrade of 100 MHz for TVB. But no change in all core boost clocks(Turbo oost max). Everything else is almost exactly the same. What a fantastic upgrade. Run and buy. Why bother put resources on an refresh? Plus models? LOL

 

 

Intel Core Ultra 290K, 270K and 250K Plus spec leak: “Arrow Lake Refresh” with higher clocks, more cores and faster memory support

 

Core Ultra 200K Plus (overview).

Core Ultra 9 285K➡️ 290K Plus: + 100 MHz P-Core (Max)

Core Ultra 7 265K➡️ 270K Plus: + 4 E-Cores

Core Ultra 5 245K➡️ 250K Plus: + 100 MHz P-Core (Max), + 4 E-Cores

Did you mean to link to a post by @electrosoft or an article about the shiny new product that offers so little? The link leads here:

https://notebooktalk.net/topic/109-official-benchmark-thread-post-it-here-or-it-didnt-happen-d/page/824/

 

[Papusan]

6 minutes ago, Mr. Fox said:

Did you mean to link to a post by @electrosoft or an article about the shiny new product that offers so little? The link leads here:

https://notebooktalk.net/topic/109-official-benchmark-thread-post-it-here-or-it-didnt-happen-d/page/824/

 

 

Nope. Not sure what happened. Fixed now 🙂

[Mr. Fox]

2 hours ago, Papusan said:

Nope. Not sure what happened. Fixed now 🙂

Intel Core Ultra 290K, 270K and 250K Plus spec leak: “Arrow Lake Refresh” with higher clocks, more cores and faster memory support

 

So, a real "nothing burger" when all is said and done. 

 

The most intelligent thing AMD has going for it is long-term socket viability. Intel needs to learn from this now and back-track on some of their recent very stupid ideas (bring back hyperthreading and go back to monolithic design without the stupid Atom cores). It is hurting them more than they know. Their ending of hyperthreading was the final nail in the coffin for me. But their constant socket revisions requiring new motherboards after a few CPU generations is a compelling reason for people to choose AMD even if they would rather stay with Intel. (I am conflicted because I can't say that I like either one now. They both have pros and cons, but both have more cons than pros.) The short socket lifespan is a very damning attribute for Intel now that they no longer operate from a place of absolute supremacy and domination. They used to expect us to put up with it as going with the territory, which was fine when they still ruled the world.

 

I'd really like to make my next build an Intel platform again. The overclocking experience (CPU and memory) is much better on Intel, but the lack of hyperthreading and abbreviated upgrade path due to frequent socket changes is making me question the logic of it. Once upon a time, not long ago, that was the only somewhat logical basis for choosing to go with AMD instead of Intel and that alone was never enough because there were too many other compromises attached to the idea.

Quote

The last breath for the LGA-1851 socket. Intel is now set to introduce its Arrow Lake Refresh next year, likely as a last push before moving to the LGA-1954 socket for Nova Lake. There is still no clear communication on the launch date for that successor. For now, Arrow Lake Refresh is shaping up as the final upgrade for the LGA-1851 platform.

Intel should revive X299 and make modern versions of CPUs for that socket LGA 2011-v2 platform... 32 or 36 threads at 6.0GHz+ all-core overclock, quad-channel DDR5 8000+ and 40+ PCIe lanes... I'd be all over that, like white on rice... hell yeah, in a heartbeat. And, I'd be willing to pay twice as much as a Core Ultra (or Ryzen) flagship CPU to have it. The most fun I have ever had with overclocking was on X299. 100% (double stock clock) overclocking wasn't that difficult.