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

[Talon]

On 9/10/2022 at 4:51 AM, jaybee83 said:

welp, thats all folks, nothing more to see here... 😔

 

 

 

https://twitter.com/JohnOspanov/status/1569233547218812928?s=20&t=QXvb2Ta35DNHRcqS1GYgUw

 

While I suppose there is always a chance, I am going to listen to the guy that is there talking to Intel directly at an event. Until Intel announces anything, it's all rumor. 

[electrosoft]

1 hour ago, Talon said:

 

https://twitter.com/JohnOspanov/status/1569233547218812928?s=20&t=QXvb2Ta35DNHRcqS1GYgUw

 

While I suppose there is always a chance, I am going to listen to the guy that is there talking to Intel directly at an event. Until Intel announces anything, it's all rumor. 

 

Like I posted earlier, MLID has been wrong numerous times in the past.....

 

 

[Papusan]

16 hours ago, Tenoroon said:

Speaking of cleaning laptop keyboards. I usually wipe down my laptops with 91% Isopropyl Alcohol, but that seems to always leave these darkish stains in the keyboard. Is there anything you guys use to clean keyboards that doesn’t “stain” them?

 

15 hours ago, Mr. Fox said:

I am not sure if I have seen an example of the stains you mention, but the issue I have always had getting a keyboard (desktop or laptop) to look like new is the fact that the keys are polished and have shiny spots from use. Nothing can be done to make it look new again on a laptop other than replace the keyboard.

Change lights settings in the house/room and the whole will look much better. And cheaper than replacing parts, LOOL

[jaybee83]

4 hours ago, Talon said:

 

https://twitter.com/JohnOspanov/status/1569233547218812928?s=20&t=QXvb2Ta35DNHRcqS1GYgUw

 

While I suppose there is always a chance, I am going to listen to the guy that is there talking to Intel directly at an event. Until Intel announces anything, it's all rumor. 

im actually rooting for him being wrong!

 

2 hours ago, electrosoft said:

 

Like I posted earlier, MLID has been wrong numerous times in the past.....

 

 

and hes also been right a lot of times 😄 

[Mr. Fox]

This fella has some hella electronics repair skills. Most people would have taken one look at it and decide it was over.

 

 

[electrosoft]

7 hours ago, jaybee83 said:

im actually rooting for him being wrong!

 

and hes also been right a lot of times 😄 

 

Such is the life of a rumor monger.... 😁

[ryan]

my best cpu score yet don't think iv'e broken 23000 before

 

[Mr. Fox]

Here is some Sabrent Rocket DDR5. More to follow on this in a week or so. Another example of Hynix M-die overclockability. Note the DRAM:FSB ratio.

 

[jaybee83]

5 hours ago, Mr. Fox said:

Here is some Sabrent Rocket DDR5. More to follow on this in a week or so. Another example of Hynix M-die overclockability. Note the DRAM:FSB ratio.

 

huh, i had no idea sabrent also does RAM! neat 🙂 are those 4800 Mhz modules?

[Mr. Fox]

2 hours ago, jaybee83 said:

huh, i had no idea sabrent also does RAM! neat 🙂 are those 4800 Mhz modules?

Yes, Sabrent is a newcomer to market with RAM. They are the same M-die chips as what is found on my TG Delta and generic OEM modules and they overclock the same.

 

It seemed odd at first, but I actually like not having any XMP profiles better. This was common on laptops, most of which couldn't benefit from XMP, but the generic Hynix sticks are the first desktop sticks I have owned with no XMP profiles since XMP was invented. Manual overclocking is always better because the XMP programming is usually half-assed, and the added cost of memory with XMP profiles added seems like a waste of money now. The Sabrent modules are my second pair without XMP. In light of how that sucks, as does the pathetic fake heat sink/no cooling garbage, on retail gamer memory trash, there is no point in leveling up with a brand name. You're actually leveling down and flushing your money down the toilet on nothing.

[Mr. Fox]

 

 

[ryan]

i dont get it I can get 22500-2300 gpu score sometimes and a max cpu of 24000 and a combined of 9000-9300 which means I can break 20k in firestrike but it never happens. I cant get my gpu running much better Ive tried almost everything the cpu is simple increase speed and lower thermals. gpu seems to bounce around with scores and speeds I have no idea why nvidia didnt just make the cards like they were with the 970m ect

[Mr. Fox]

12 minutes ago, ryan said:

i dont get it I can get 22500-2300 gpu score sometimes and a max cpu of 24000 and a combined of 9000-9300 which means I can break 20k in firestrike but it never happens. I cant get my gpu running much better Ive tried almost everything the cpu is simple increase speed and lower thermals. gpu seems to bounce around with scores and speeds I have no idea why nvidia didnt just make the cards like they were with the 970m ect

Buggy software. I have seen overstated Fire Strike scores that were totally impossible to achieve or replicate using the hardware they are run on, but validated by 3DMark due to glitches in the benchmark that skewed the results.

 

Sometimes the overall score makes no sense, but it is calculated based on an algorithm that gives the most weight to one of the component sub-scores and ignores or discounts stellar performance to another component. Fire Strike is guilty of that. You can get a better overall score on a less powerful system with a strong GPU because Fire Strike doesn't give much credit for having a monster CPU. The combined score can bring down the overall score. You can have an insanely good graphics score and a sucky overall score if the combined test glitches out, and the combined test often glitches and can vary widely by GPU driver version.

 

A laptop that "balances" power draw when the CPU and GPU are both under heavy load will commonly suffer in the combined test, which makes a big hit on the overall score.

[ryan]

yeah like in my mind im like ok im going to break 20k thats all I want from this system but it never happens one score drops while another rises. also I love you guys best people on the interent

[jaybee83]

8 hours ago, Mr. Fox said:

Yes, Sabrent is a newcomer to market with RAM. They are the same M-die chips as what is found on my TG Delta and generic OEM modules and they overclock the same.

 

It seemed odd at first, but I actually like not having any XMP profiles better. This was common on laptops, most of which couldn't benefit from XMP, but the generic Hynix sticks are the first desktop sticks I have owned with no XMP profiles since XMP was invented. Manual overclocking is always better because the XMP programming is usually half-assed, and the added cost of memory with XMP profiles added seems like a waste of money now. The Sabrent modules are my second pair without XMP. In light of how that sucks, as does the pathetic fake heat sink/no cooling garbage, on retail gamer memory trash, there is no point in leveling up with a brand name. You're actually leveling down and flushing your money down the toilet on nothing.

ha, im only reporting from laptop-land still but ive never used any xmp profiles cuz they never booted in the first place *lol* so it was always manual RAM tuning for me 😄 

[Papusan]

11 hours ago, ryan said:

i dont get it I can get 22500-2300 gpu score sometimes and a max cpu of 24000 and a combined of 9000-9300 which means I can break 20k in firestrike but it never happens. I cant get my gpu running much better Ive tried almost everything the cpu is simple increase speed and lower thermals. gpu seems to bounce around with scores and speeds I have no idea why nvidia didnt just make the cards like they were with the 970m ect

 

11 hours ago, ryan said:

yeah like in my mind im like ok im going to break 20k thats all I want from this system but it never happens one score drops while another rises. also I love you guys best people on the interent

3DMARK FIRESTRIKE total score is mostly about graphics score. Rather cripple the Cpu speed and put more works into the graphics (run the GPU at lowest possible temp and give it most of the power headroom). Give a huge thanks to Nvidia and the laptop OEMs for destroying own products. And a good advice... Use NVCleanstall for fast driver installs and test out different drivers for your graphics card. Some drivers will give you a slightly better results. A lot drivers to test for your card. Test and test bro Ryan. Don't give up. And only install the bare minimum from the graphics driver package. For the testing you don't even not need to run DDU in between the different drivers and tests. Only do that for final install.

 

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

 

This is just disgusting.... 4080 with two different die's. 

 

 

https://videocardz.com/newz/nvidia-geforce-rtx-4090-rtx-4080-16gb-12gb-max-tgp-and-gpu-clocks-specs-have-been-leaked

 

And yep.... Nvidia will give almost the same love for the 4090 as they did for 3090.

 

Edit. Maybe Nvidia and AMD still have some hopes. Or maybe they have invested some money to help the miners to keep up maximum profits…

 

https://wccftech.com/bad-news-gamers-ethereum-mining-is-not-going-away-not-yet-at-least/

 

When it rains on the priest, it drips on the bell ringer.

https://no.m.wiktionary.org/wiki/når_det_regner_på_presten,_drypper_det_på_klokkeren

[jaybee83]

17 minutes ago, Papusan said:

This is just disgusting.... 4080 with two different die's. 

 

 

https://videocardz.com/newz/nvidia-geforce-rtx-4090-rtx-4080-16gb-12gb-max-tgp-and-gpu-clocks-specs-have-been-leaked

 

And yep.... Nvidia will give almost the same love for the 4090 as they did for 3090.

 

wth man...Nvidia doing their best to confuse their customers "oh look, that other 4080 only has 4 gigs less vRAM but so much cheaper! same card, same performance! such a DEAL!" 🤦‍♂️☠️

 

on the bright side, just received my Samsung Odyssey Neo G8, holy crap is that a HUGE monitor! and amazing colors and blacks, dayum... still need to do some finetuning though. plus, my laptop isnt even strong enough to drive it at 240 Hz, max i can get is 96 🤣 just a fewwww more weeks...

[Talon]

On 9/8/2022 at 11:20 PM, Mr. Fox said:

 

 

 

Well... as luck would have it, these naked generic DDR5-4800 SK Hynix modules with no XMP that I purchased brand new on eBay for $80 each overclock better than my TeamGroup RGB sticks, LOL. And, totally naked with nothing but a fan blowing on them they run a good 15-20°C  cooler than memory with stock trash heat sinks beautification covers on them. What a wonderful find at an affordable price. Using exactly the same manual settings they perform better.

 

@johnksss

 

This is actually as cool as the stick were running on water. It seems like the stock heat sinks we are getting on the overpriced brand-name memory are actually trapping heat rather than shedding it.

 

Conclusion: If you are looking for performance rather than aesthetics, then don't waste your hard-earned money on branded sticks with worthless beautification covers inappropriately labeled as "heat sinks" LOL. $80 each, brand new... what's not to love? And, no need to fart around with buggy nightmare RGB trash software messes just to kill an RGB rainbow puke parade.

 

Next Steps: Back to the water cooling so I can see how far these sticks can go.

 

Good stuff! I am hoping that once I get my green PCB A-Die or whatever I end up getting I can really push D5. I saw a screenshot today with 6600 Team Group A-Die that was at 7600 on a Z790 4 Dimm Hero board with a 13900K. 

 

I think I finally stabilized 7000 CL32. Looking like all I needed was to increase my VCCSA a tad to a +90mV offset. Just passed 5000% and no issues. 

[Papusan]

4 hours ago, jaybee83 said:

 

wth man...Nvidia doing their best to confuse their customers "oh look, that other 4080 only has 4 gigs less vRAM but so much cheaper! same card, same performance! such a DEAL!" 🤦‍♂️☠️

 

on the bright side, just received my Samsung Odyssey Neo G8, holy crap is that a HUGE monitor! and amazing colors and blacks, dayum... still need to do some finetuning though. plus, my laptop isnt even strong enough to drive it at 240 Hz, max i can get is 96 🤣 just a fewwww more weeks...

Same performance… Nope. Less cores as well on the Max-Q version. And slower vRam and lower TGP. Ala fully castrated.
 

Put it the other way… Desktop cards as the weakest 4080 of those two should be branded Max-Q in same way as Nvidia did nfor the older gen Mobile graphics cards. Nvidia finally gave the desktop buyers same nice graphics options as laptop-jockey’s now. And I don’t joke about this🙂 Isn’t it nice… That Nvidia will treat all types gamers exactly the same way?

[ryan]

so iv'e been following this thread for a while now! learned alot...I bought this ram and  decided to overclock. all kidding aside this is me just imagine a colt 45 in one hand and a hammer in the other lol.

 

 

[Papusan]

More pict of the crocks bro @electrosoft🙃

 

What a heavy baby. Slightly below 1.3 kg. And the Alphacool XPX block clocking in at wimpy 0.350 kg. And this with an heavier steel backplate for the XPS block vs the lighter alu backplate from EK. 

 

 

But what a stupid block design from EK. You need to be born as an Vishnu or being married. Two hands ain't enough to mount this fully metal baby in your chassis. This block is meant to be mounted on the MB before you put it into your chassis. Yep, damn stupid design from EK.

[KING19]

10 hours ago, ryan said:

yeah like in my mind im like ok im going to break 20k thats all I want from this system but it never happens one score drops while another rises. also I love you guys best people on the interent

 

Firestrike is pretty outdated these days especially with Modern CPUs and GPUs. Timespy scoring system is a improvement over firestrike.

 

I was bored so i decided to run it and i almost broke the record score lol

 

Mind well post this too i ran timespy almost a week ago and my score increased for some reason. My scores is usually around 6300 with the overclock but end up getting 6408 my highest score ever:

 

Havent updated my GPU drivers since my last windows reinstall almost a year ago.

[Tenoroon]

I decided to ditch the old thick thermal pads in favor of some K5 Pro for my 1080 in the P775TM. I did a bit of a heatsink mod to make sure the heatsink can make as much contact with the die and VRM as possible (not paperclip mod.) There wasn't much of a temp improvement surprisingly, but I'm a bit worried as I had to use a lot of K5 Pro and while it makes decent contact, theres still quite a thick layer of K5 between the heatsink and power delivery/VRAM of the card. Would using shorter screws or the paper clip mod still help me increase the mounting pressure?

[jaybee83]

5 hours ago, KING19 said:

 

Firestrike is pretty outdated these days especially with Modern CPUs and GPUs. Timespy scoring system is a improvement over firestrike.

 

I was bored so i decided to run it and i almost broke the record score lol

 

Mind well post this too i ran timespy almost a week ago and my score increased for some reason. My scores is usually around 6300 with the overclock but end up getting 6408 my highest score ever:

 

Havent updated my GPU drivers since my last windows reinstall almost a year ago.

ha, just goes to show that overclocking and score chasing is also just a bit of luck of the draw. if the stars, the moon and ur hardware align u might just get those few extra points to be the record. i expect follow-ups to get that score above 14927! 😉 

 

5 hours ago, Tenoroon said:

I decided to ditch the old thick thermal pads in favor of some K5 Pro for my 1080 in the P775TM. I did a bit of a heatsink mod to make sure the heatsink can make as much contact with the die and VRM as possible (not paperclip mod.) There wasn't much of a temp improvement surprisingly, but I'm a bit worried as I had to use a lot of K5 Pro and while it makes decent contact, theres still quite a thick layer of K5 between the heatsink and power delivery/VRAM of the card. Would using shorter screws or the paper clip mod still help me increase the mounting pressure?

id say theoretically yes, every little bit of distance would help improve that thermal conductivity gradient through the applied K5 Pro layer.

[Mr. Fox]

8 hours ago, Papusan said:

More pict of the crocks bro @electrosoft🙃

 

What a heavy baby. Slightly below 1.3 kg. And the Alphacool XPX block clocking in at wimpy 0.350 kg. And this with an heavier steel backplate for the XPS block vs the lighter alu backplate from EK. 

 

 

But what a stupid block design from EK. You need to be born as an Vishnu or being married. Two hands ain't enough to mount this fully metal baby it in your chassis. This block is meant to be mounted on the MB before you put it into your chassis. Yep, damn stupid design.

I find the mounting arrangement very inconvenient. Accessing the back of the motherboard while it is installed is easy on either of my cases but I know for some people that may not be true. Some motherboard tray are poorly designed and in some cases other things get in the way. But it is not convenient having to access both sides to remove the water block even when access is not impeded by other things. That being said, it looks good and the temperatures are very good. Then convenience is less problematic given that I'm using it on my work computer and I'm not frequently taking it apart and experimenting with different things like I am my benching rig.

 

Did it help temperatures in your case?

8 hours ago, KING19 said:

 

Firestrike is pretty outdated these days especially with Modern CPUs and GPUs. Timespy scoring system is a improvement over firestrike.

 

I was bored so i decided to run it and i almost broke the record score lol

 

Mind well post this too i ran timespy almost a week ago and my score increased for some reason. My scores is usually around 6300 with the overclock but end up getting 6408 my highest score ever:

 

Havent updated my GPU drivers since my last windows reinstall almost a year ago.

Same answer as Fire Strike. Buggy software and spastic laptops. Swings in scores and inconsistent results are expected and normal.

8 hours ago, Tenoroon said:

I decided to ditch the old thick thermal pads in favor of some K5 Pro for my 1080 in the P775TM. I did a bit of a heatsink mod to make sure the heatsink can make as much contact with the die and VRM as possible (not paperclip mod.) There wasn't much of a temp improvement surprisingly, but I'm a bit worried as I had to use a lot of K5 Pro and while it makes decent contact, theres still quite a thick layer of K5 between the heatsink and power delivery/VRAM of the card. Would using shorter screws or the paper clip mod still help me increase the mounting pressure?

K5 Pro offers no resistance to compression when used in place of a thermal pad, so it will not interfere with contact in any way. As long as you apply enough that it squishes out between the contact surfaces you'll be good to go, and the contact pressure would be the same or better (expect better) at the die when using K5 Pro versus thermal pads. K5 Pro is not better than pads as a TIM, and in some cases not as thermally conductive, but using it eliminates the possibility that thermal pads are interfering with contact and it maximizes contact on the surfaces where it is applied. That is the reason it is a good option. Given the haphazard and sloppy manner in which laptops are manufactured it covers a multitude of sins.

 

Using shorter screws won't change anything. On the GPU you tighten the screws until the tension arm bottoms out against the X bracket. It's not going to get any tighter than that unless you increase resistance to tightening the screw by making the tension arm resist being pressed down against this standoff on the X bracket. Getting rid of the springs on the CPU and using ordinary screws will let you increase contact pressure on the CPU die. You have to pay attention to what is happening and not break anything, but you'd have to be kind of reckless to create a problem with over-tightening.