Bastiaan_NL

Some extra information from @leeghoofd:

"Only Microsoft operating Systems allowed
BenchMate 10.12.2 version only
CPU-Z 2.06 or newer unless specifically mentioned
GPU-Z 2.54
All other benchmark versions need to be updated including eg the Systeminfo 2.64.1188 or newer
Monitoring, systeminfo remain enabled for all UL Benchmarks"
And as always, check the specific rules for the benchmark you are running on the HWBOT page before submitting. If you have doubts, DM the team captains.

4 hours ago, Bones said:

BTW in case you guys missed it, for AM3+ I was able to get the bot to allow a couple of 4 core chips for the 7-Zip part, namely FX-4100 and FX-4300 chips.
I reminded them there are only a few boards (Mostly Gigabyte boards) that can truly support 1 core/1 module operation (4P) and because of this restriction there woudn't be too many with subs in that category - Effectively making it more or less a "Giga-Fest" instead of a truely competitive stage.

I did follow the conversation where you mentioned this, and I'm glad you did. An FX4300 is on the way to me to make it easy

With Trixx it's hit or miss, at least that's what I hear on the hwbot discord. I never tried it myself.
Elmor is working on the controller of my Toxic, let's hope he can get it to work in time. He bought a card which should have the same controller as mine, but sadly it was different.
1.3v is good for 1070mhz core in Heaven, after that it's a nice disco   

This one was actually caused by memory, but too funny not to share:

No LN2 plans yet, but I'm working on GPU brackets for the cascade with a coworker. 3D printing the designs and laser/water cutting them from metal after test fitting.

Also, I've got a GPU pot for dice/LN2 if needed.

1 hour ago, Sir Beregond said:

I like how the box for that first card says it is designed for extreme graphics vs say regular graphics. 

Yeah, I hope it's also designed for extreme cooling  

The first few victims showed up for the team cup

HEAR YE HEAR YE!!

We are quickly closing in on the start of the 2023 HWBOT Team Cup. 
The preparations have begun, some of us are already searching the local marketplaces for deals on hardware we could use to earn our spot amongst the big teams.
If you find yourself in the posession of hardware listed in the spreadsheet, you are very much welcome to join us. Please DM me on Discord if you want to be added to the spreadsheet or if you have a lot of the listed hardware I can give you permissions to edit the sheet.
The competition will start 07-15-2023 at noon (need to figure out the time zone) and we'll try to get a countdown timer up.

Some extra information from @leeghoofd:
"Only Microsoft operating Systems allowed
BenchMate 10.12.2 version only
CPU-Z 2.06 or newer unless specifically mentioned
GPU-Z 2.54
All other benchmark versions need to be updated including eg the Systeminfo 2.64.1188 or newer
Monitoring, systeminfo remain enabled for all UL Benchmarks"
And as always, check the specific rules for the benchmark you are running on the HWBOT page before submitting. If you have doubts, DM the team captains.

Sadly it's rather expensive to get the EHW merch here. 
But the good news is, @Fluxmaven gave me a pair of the EHW stickers he printed. 
The new bench room is still a work in progress, but once that's all taking shape I want to have those somewhere in sight for submission pictures or streams

8 hours ago, J7SC_Orion said:

 

...can't quite tell the fan air flow direction (looks like blowing inwards towards the socket though)... fyi, some XOCers like the two helper fans turned around so that they help evacuate potentially moist air away from the socket and surrounding electrical bits...in my personal experience though, it doesn't make a ton of difference

I expected you to comment on that

The fans are blowing inward.

So I see little to no ice buildup around the socket after a 2 hour session. The board is covered with liquid tape between the inductors and the same area on the back. The back has one layer of armaflex clamped between the board and the LN2 pot backplate and a second layer underneath and there is an armaflex ring around the evap base. I stuff as much shop towel between the threads of the mount and the evap base/armaflex to catch whatever moisture escapes, but after breaking down they were all dry. It feels very efficient, for what I can tell with limited experience. The fans are there to cool the VRM, but if I notice any moisture I will add extra fans in an outward direction. As for LN2, I bet that will make a bit more difference with buildup.

1 hour ago, damric said:

lol if it was that cool here then I'd be breaking the records

Glad I'm not over there

Once it cools down in the house again I will try to do some more decent runs on mine.
It's been over 82f/28c for a week in here, not a great starting point for benchmarks

An overview of the setup, 12600K, Z790 Apex and G.Skill A-die:

And a webcam shot of the setup, notice the temperature drop during a run and the temperature climbing after the run:

For a while I've been the owner of a cascade phasechange unit. I have used it a couple of times for testing, and recently did some benchmarks with the cooler.
I'm trying to figure it all out, and given the lack of time recently there isn't a lot of progress but I'm trying to record some of it so I can share that here or elsewhere. 
There are more ideas brewing, but I'm not rushing any of that. More to come, hopefully  

7 hours ago, damric said:

I might have to rerun these. i saw in the screenshots that the guy in first place is using a sort of LOD utility mod. I think it's allowed, too.

It is allowed and the guy in first knows what he's doing. I expect a lot more of those runs, a few on cascades and maybe dice. Not sure if LN2 will make a difference, they were talking about negative scaling with mem frequency on these.

Sorry for my absence here, demanding project at work and rising temperatures keep me from doing anything productive.
It seems like there are some tweaks which actually work but I haven't tried any of it. As for hard modding, on most cards it's doable. Maybe @Avacado wants to mod your card @damric because I modded his card
The controller on my Toxic is still not supported by the EVC2, so that one is not ready yet. The Club3d card has shitty mem but at least the voltmods are working. I've got some baseline results but didn't bother to upload them right at the start. Maybe I should just freeze that card right away

22 hours ago, Avacado said:

It's not a true CL15. Used EVGA ELEET Software to modify timings from within windows. None of those settings were bios set. Honestly impossible to really know what the CL was. 

Afaik all (the ones I tested) timings can be changed in windows besides CL. CL needs training, and CL15 7000 is for sure a nogo

21 hours ago, kaliz said:

Awesome stuff, do you happen to know a method for my msi b760 mortar max board to use that software? Or anything that can OC in windows?

I'd try the HWBOT OCsoftware section on Discord, usually that's your best bet. Given that it's a b series board I doubt it supports OC software.

4 hours ago, Avacado said:

It sounds like you have already done your homework. I doubt EVGA ELEET will work with your board but you can try. I think @Bastiaan_NL might be a better resource than me for this topic. 

Besides the two XOC boards I have my software knowledge is limited. 

5 hours ago, damric said:

Wow looks like you guys were busy working hard. White board checked off? 

 

I'd like to know the results of those volt mods

The whiteboard is checked off, the nice thing is you can just erase whatever you can't finish.

So most of the GPU's have only been tested briefly for functionality but have not been benched besides a quick 3dmark vantage performance run. @Avacadowill eventually have to do his cards, @Fluxmaven has a modded HD6850 now and @firedflyhas a modded GTX680, Quadro 600 and a pair of GT210's to mod himself.

10 hours ago, J7SC_Orion said:

Is it time yet to unleash the 'There goes the neighborhood' ? 

 

But seriously folks, wish I could be there, but duty (such as arguing with several gov't departments here at the same time) calls. 

 

Have fun, y'all !!

Thanks J7, it's a lot of fun

First modded GPU over here, HD6850 for Chris

IN

You forgot to mention a few things on the to do list which I probably shouldn't mention here either, but it's probably good for the team spirit

22 hours ago, Kaz said:

Are there any chips that can utilize that ram speed?

2 hours ago, bonami2 said:

@Bastiaan_NL had luck running higher than that if am not wrong on a kingpin or apex if am not wrong.

 

Depends on what I'm doing, but for now I'm doing fine at 8000 CL32 on the Z790 APEX with a 13900KS. Some guys are benching 9000mhz on air, not sure about the timings etc but there are some crazy results out there. They are in a voltage range on some settings where I prefer to stay away from, if you want your hardware to last a little longer than a month or two

20 minutes ago, J7SC_Orion said:

 

...that never stopped me before    just think, a whole wide field of  memory rabbit holes to dive into 

Good luck breaking your legs.. 
I've got enough fun with getting two sticks to work properly

12 hours ago, Mr. Fox said:

Hey bro. I have the 6900 XT. Do you or @pioneerisloudknow of any AMD driver mods or tool to mod the driver that gives you basic control panel access without the worthless payload of garbage?

 

I use NVCleanStall to strip out all of the useless trash that the Green Goblin packs into the GeFarts drivers (like GeFarts Experience and what not). I'm really not liking this gamerboy rubbish included in the Adrenaline "Pro" malware.

11 hours ago, Mr. Fox said:

Yeah, that gosh-awful overlay thing... talking about being OD'd on stupid... makes me want to punch somebody... real bad.

You have no idea how much I love reading your rants about things like this, as I can relate. I just hate any software where they invested more time into making it look good rather than making it work properly.
Give me software like the ASUS APEX Tool or HOF_NVVDD, not some useless fancy looking AMD or iQue software

Heck, I wouldn't mind if my windows looked like XP and would be as lightweight with just the options I need for it to function. No, instead I'm spending hours on stripping an installation for it to not use loads of memory right after installation with all the crap they force down our throats..

(opened it on the main for Screeny)

/rant 

Have fun with the 6900XT @Mr. Fox, I'm still waiting for them to become real cheap and have some fun on cold water maybe
 

13 minutes ago, J7SC_Orion said:

  

...not sure whether clocks are 'unrealistic' per se (my 3090 Strix has been at 2280 or > for all these years). Key is to remember that this bench and its sub-components is meant for cross-platform, so power limited. The peak watts for all my GPUs is low with Geekbench-6, thus they don't heat up internally as much (up to 76.3 Billion transistors on N4 ) and can therefore hit their best levels.

 

FYI, I have used the 3240 MHz curve of the RTX 4090 for Superposition 8K at > 560W so way more peak watts; I probably could add some more clocks for Geekbench-6 ...will test that out next weekend, along with EEC setting. In any case, similar GPU types face the same benefits / limitations with Geekbench-6 so not a skew.

 

 

 

---

 

Regarding the Elmorlabs RTX VRAM heater, I know someone who uses it with a chiller setup and it apparently it works great. But that's also another reason I don't do HWBot anymore - I still like a bit of benching now and then, but then turn around w/o rebooting and do some productivity / work on it, then play some FS2020.  Same cooling setup...one-size-fits-all settings, apart from MSI AB clock profiles 

 

 

I know what you mean, but if I would keep those settings and run Port Royal next it'll crash while collecting the system info. 
You could say that the driver isn't really optimized for those less demanding tasks, where it could actually boost higher and possibly keep those clocks a bit longer instead of downclocking in between. Precision X1 has the nice 'boost lock' function, but if I use X1 with the current bios on my card it'll try to 'update' the firmware. Like I said before, it is a fun benchmark to try in between but I would never use this to compare cards myself. 

So I ran GB6 on the air cooled 3090 FTW3 with the following result:

As fun as this may be, I think that this is another benchmark where you can use unrealistic clocks (same as GPUPI).
I was running the core at 2280-2295mhz though the actual clocks under decent load from a 3D benchmark are between 2160 and 2200mhz. The same goes for the memory, it was scaling all the way up to +900mhz where I usually sit around +750 for an artifact free image. 

@J7SC_Orion Regarding the heater, I'm really curious how much difference it would make on water.
It'll also dump extra heat into the water, which will have a negative effect on the core temperature. For anything colder on the core itself a heater like that is really useful, and probably needed to compete

4 hours ago, J7SC_Orion said:

   

...wasn't meant as a serious tease other than responding to @bonami2 re. more recent AMD models' speed restrictions vs. MPT. Besides, the result is obviously posted at Geekbench, ditto for the 6900XT, and the other cards.

 

I also tagged you and @Bastiaan_NL because many HWBotters run these new cards with GDDR6X 2GB chips too cold...that specific VRAM type (unlike its 1 GB cousin) likes at least 55 C to reach its max MHz, yet the VRAM sits much closer to the core than in other models. Said VRAM also cools down very quickly after a second or two at idle.

 

--- 

 

This may sound odd, but I actually don't own a smartphone (anymore) - I am already online for too many hours...still Geekbench 6 is cross-platform and also works for Apple iPhone and Android smart phones, as well as tablets etc. It would be great if we can get some Geekbench 6 compute results for those as well 

 

 

 

We've been talking about that temperature subject on the memory chips. The problem is that you want to cool the VRM and core but want to run the memory at a stable ~55°C to get the most out of it. My blocks are either full cover or core only, so it's always a compromise trying to get the best performance.