• Welcome to Andy's Workshop Forums. Please login or sign up.
 

Hacking an HP Z800 motherboard into a standard PC case

Started by Andy Brown, November 01, 2014, 04:53:46 am

Previous topic - Next topic

LordOdin

^ Actually there is a little dip behind the fan mounts that you cant see in these images. Here is the best example I could find lol its a pretty bad example you will have to use a little imagination.


JonMS

Quote from: Vesuveus on July 11, 2015, 06:36:42 am
Pretty sure the hyperthreading issue was in windows 10. It may have been fixed with an update. Definitely not related to hardware.

I ran the early tech previews with no problems at all.

Attilio Fiandrotti

Hello, not sure if this topic has already been discussed before, but has anyone attempted to get correct ambient temperature readings so far ? It seems that the HP zx00 and xwxx00 (at least some) workstations read the ambient temperature via a sensor connected to pins 11-13 of the main connector. The sensor is embedded into the power cable assembly, at least for the xw8600 and similars (P/N: 349576-006). Here [1] is reported an equivalent schematic to build your own sensing circuit. It seems that the air temp sensor drives the speed at which the front fans spin, at least on xwxx00 workstation: getting correct temp readings should slow em without fancontrol. Another option would be getting a xw8600 cable and using the sensor contained therein.

Attilio

[1] http://h30499.www3.hp.com/t5/Workstations-z-series-xw-series/xw4400-motherboard-front-control-panel-P5-Schematic/td-p/1139459?notmigrated#.Vaegqf7sxt8

LordOdin

Got it up and running with non ecc ddr3, works flawlessly and the 2 x5570s put my FX 8350 to shame :D

Here is a little chart im working on comparing render times in Blenders render engine Cycles

https://docs.google.com/spreadsheets/d/1EBtZlioApDvo6vmRWeFbE2XF3A6B_9b7pKLMqWru3TA/pubhtml

Marco Silva

Hi, your table doesn't seem too fair to me.

You shouldn't be comparing used cpu prices with brand new prices, it seems in an exaggerated way that you have really good value for the money.

Also i have an overclocked 4930K as my daily driver and it's faster than my dual X5675 in blender cycles, so your numbers seem way off to me...


Andy Brown

Quote from: Attilio Fiandrotti on July 16, 2015, 06:31:07 am
Hello, not sure if this topic has already been discussed before, but has anyone attempted to get correct ambient temperature readings so far ? It seems that the HP zx00 and xwxx00 (at least some) workstations read the ambient temperature via a sensor connected to pins 11-13 of the main connector. The sensor is embedded into the power cable assembly, at least for the xw8600 and similars (P/N: 349576-006). Here [1] is reported an equivalent schematic to build your own sensing circuit. It seems that the air temp sensor drives the speed at which the front fans spin, at least on xwxx00 workstation: getting correct temp readings should slow em without fancontrol. Another option would be getting a xw8600 cable and using the sensor contained therein.


I'm going to try building one of these. I've got a basket of parts that I need to order from Farnell within the next few weeks. I'll probably use the exact same MMBT3904 part in the SOT-23 surface mount package rather than hunting around for a through-hole transistor with the same characteristics.
It's worse than that, it's physics Jim!

Attilio Fiandrotti

Quote from: Andy Brown on July 16, 2015, 01:21:01 pm
I'm going to try building one of these. I've got a basket of parts that I need to order from Farnell within the next few weeks. I'll probably use the exact same MMBT3904 part in the SOT-23 surface mount package rather than hunting around for a through-hole transistor with the same characteristics.


Great, looking forward to the results of your experiment. As a poor man's solution to the subproblem of slowing down the front fans, maybe one could fool the logic controlling the fan speed by hooking up a simple resistor with proper impedence (assuming the logic reads on the current flowing through the sensor) ?

Attilio

LordOdin

Quote from: Marco Silva on July 16, 2015, 12:54:50 pm
Hi, your table doesn't seem too fair to me.

You shouldn't be comparing used cpu prices with brand new prices, it seems in an exaggerated way that you have really good value for the money.

Also i have an overclocked 4930K as my daily driver and it's faster than my dual X5675 in blender cycles, so your numbers seem way off to me...


Im just reporting the times my friends told to me. The price part was added as a after though and I just did the prices from my memory.

But if there is something I know more about than computer hardware its Blender and Cycles.

https://mega.co.nz/#!fdonGKwI!6IBcEPDqmwjDaCzjfp7BpdjTUTmZj1vOy-jaC5-T8VA

Here are the scenes. If you want to have a go I could add you to the chart. I am using 2.75, dont change any settings just hit f12 tile sizes and all are set up already.

BenGman

Quote from: Marco Silva on July 16, 2015, 12:54:50 pm
Hi, your table doesn't seem too fair to me.

You shouldn't be comparing used cpu prices with brand new prices, it seems in an exaggerated way that you have really good value for the money.

Also i have an overclocked 4930K as my daily driver and it's faster than my dual X5675 in blender cycles, so your numbers seem way off to me...


I think it's good to have the used prices because I would see if it's cost effective to buy used components over new/recent CPUs based on performance once the used components are available. I would have never started this project if I knew I was going to purchase new CPUs even though I bought my Z800 v3 board brand new.

Marco Silva

July 19, 2015, 06:23:26 am #174 Last Edit: July 19, 2015, 11:09:56 am by Marco Silva
Quote from: LordOdin on July 18, 2015, 09:18:28 am
Quote from: Marco Silva on July 16, 2015, 12:54:50 pm
Hi, your table doesn't seem too fair to me.

You shouldn't be comparing used cpu prices with brand new prices, it seems in an exaggerated way that you have really good value for the money.

Also i have an overclocked 4930K as my daily driver and it's faster than my dual X5675 in blender cycles, so your numbers seem way off to me...


Im just reporting the times my friends told to me. The price part was added as a after though and I just did the prices from my memory.

But if there is something I know more about than computer hardware its Blender and Cycles.

https://mega.co.nz/#!fdonGKwI!6IBcEPDqmwjDaCzjfp7BpdjTUTmZj1vOy-jaC5-T8VA

Here are the scenes. If you want to have a go I could add you to the chart. I am using 2.75, dont change any settings just hit f12 tile sizes and all are set up already.


Ok no problem, but the operating system also plays a big role with blender...it's faster with linux for cpu rendering.

I've got an older X58 system with a X5650@4GHz dual booting to see the differences between OS's

BMW

W10 - 45.39S  Ubuntu - 36.23S

CORNELL

W10 - 26.19S  Ubuntu 19.65S

HALLWAY

W10 - 2:46.21   Ubuntu - 2:16.00

My 4930K @4.5GHz running W8.1 did on the BMW scene 37.64S, on the CORNELL scene 21.14S and on the WALLWAY scene 2:10.37.

I decided to give it a go with the Z800 board and dual X5675 on linux and damn it was faster that i thought, need to check again under windows.

BMW scene 26.44S, on the CORNELL scene 10.07S and on the WALLWAY scene 1:45.17



LordOdin

July 19, 2015, 03:41:46 pm #175 Last Edit: July 23, 2015, 05:24:58 pm by LordOdin
Quote from: Marco Silva on July 19, 2015, 06:23:26 am
Quote from: LordOdin on July 18, 2015, 09:18:28 am
Quote from: Marco Silva on July 16, 2015, 12:54:50 pm
Hi, your table doesn't seem too fair to me.

You shouldn't be comparing used cpu prices with brand new prices, it seems in an exaggerated way that you have really good value for the money.

Also i have an overclocked 4930K as my daily driver and it's faster than my dual X5675 in blender cycles, so your numbers seem way off to me...


Im just reporting the times my friends told to me. The price part was added as a after though and I just did the prices from my memory.

But if there is something I know more about than computer hardware its Blender and Cycles.

https://mega.co.nz/#!fdonGKwI!6IBcEPDqmwjDaCzjfp7BpdjTUTmZj1vOy-jaC5-T8VA

Here are the scenes. If you want to have a go I could add you to the chart. I am using 2.75, dont change any settings just hit f12 tile sizes and all are set up already.


Ok no problem, but the operating system also plays a big role with blender...it's faster with linux for cpu rendering.

I've got an older X58 system with a X5650@4GHz dual booting to see the differences between OS's

BMW

W10 - 45.39S  Ubuntu - 36.23S

CORNELL

W10 - 26.19S  Ubuntu 19.65S

HALLWAY

W10 - 2:46.21   Ubuntu - 2:16.00

My 4930K @4.5GHz running W8.1 did on the BMW scene 37.64S, on the CORNELL scene 21.14S and on the WALLWAY scene 2:10.37.

I decided to give it a go with the Z800 board and dual X5675 on linux and damn it was faster that i thought, need to check again under windows.

BMW scene 26.44S, on the CORNELL scene 10.07S and on the WALLWAY scene 1:45.17



MinGW builds on Windows are near Linux speeds... Im not sure why they don't compile it with MinGW by default.

Ok so your tests, The ones that were done on 2.69 (the Z800) are completely destroyed im not even sure if we can consider those times seeing what the images look like but yeah windows times would be nice so we can directly compare it with all the others which are also on Windows :D

Nice times BTW on all the machines!

Marco Silva

I'm not sure what's going on...but  even on Windows the Z800 completely destroyed my X79 rig  8)

BMW 28.66S  CORNELL 15.98S HALLWAY 1:41.79

I would be really interested to see other results from fellow Z800 hacker's  :)



LordOdin

Quote from: Marco Silva on July 20, 2015, 12:24:17 pm
I'm not sure what's going on...but  even on Windows the Z800 completely destroyed my X79 rig  8)

BMW 28.66S  CORNELL 15.98S HALLWAY 1:41.79

I would be really interested to see other results from fellow Z800 hacker's  :)



Would be cool :D

LordOdin

July 23, 2015, 05:18:34 pm #178 Last Edit: July 23, 2015, 05:28:55 pm by LordOdin
My stuff is idling at super high temps. I plugged the 4 pin connector of the Hyper T4s in to the CPU fan headers and they turn on they are at super low RPMs though, Im sure I have missed a solution to the problem mentioned in the forum somewhere but I couldn't find one after 2 quick looks :P


Attilio Fiandrotti

Same issue here: similar RPMs and same temps. My understanding is that  by default the two CPU fans PWM controller makes the fans spin at about 1/2 and 1/3 of the fan top RPMs respectively, where the fan top RPMs depends on the specific fan model you use. So, it seems to me there are two options i) you install fans rated at about 4000 RPMs  ii) you tweak the PWM controller to make the fans spins at a higher rate using fancontrol or so. Any other idea ?