Dell Workstation Owners Club

[Oldschool297]

Which is the best 6 core processor for T3500?

 

[lexluthermiester]

Which is the best 6 core processor for T3500?

Prices are all really low right now, so go for a W3680 oe X5680. If you have to go cheaper, a W3670 or X5675 is the next ones I'd send you toward. Make sure you update your BIOS before installing any of those.

 

[uco73]

Prices are all really low right now, so go for a W3680 oe X5680. If you have to go cheaper, a W3670 or X5675 is the next ones I'd send you toward. Make sure you update your BIOS before installing any of those.

Take X5680 or X5675. They can use up to 288GB RAM. W3680 and W3670 use up to 24GB RAM.

 

[lexluthermiester]

Take X5680 or X5675. They can use up to 288GB RAM. W3680 and W3670 use up to 24GB RAM.

I don't see the user above(nor most users) caring about that difference. Also, in a T3500 with the latest BIOS update, the limit is 48GB(6x8GB) for the W36XX CPUs. I have personally tested this.

 

[Retrorockit]

I think the actual limit for X58 is 48GB ( from posts at Overclock.Net) The W3680/3690 have an unlocked multiplier/Voltage and can be overclocked in a T3500 if that applies to your purpose.
Dell only lists the RAM capacity based on modules they tested when the computer was new. Over time larger modules become available. So the chipset/IMC limit is what actually applies in most cases..

 

[M00N]

Hello all. I know this is a bit of a stretch of a question, but i was curious as to whether i could possibly cram a T7910 mobo into my t7600 case. I got a killer deal on the 7600 and have the dual 2687ws with 256gb of ram, but this thing lacks modern cpu instructions amongst other things i have to do to get it to work properly. The dimensions almost look to line up but im not 100% sure. Any help would be appreciated, thank you

 

[lexluthermiester]

Hello all. I know this is a bit of a stretch of a question, but i was curious as to whether i could possibly cram a T7910 mobo into my t7600 case.

Simple answer is not likely. While they look like they'd fit, the T7600 motherboard and the T7910 motherboard have several differences that make swapping the T7910 mobo into the T7600 case very difficult. It's mostly a connector problem. You can try it, but be careful and don't expect it to work.

 

[Retrorockit]

Many aftermarket computers follow the ATX standard and parts swapping is normal there. But Dell builds whole discrete systems and parts interchange is very rare. Proprietary PSU formats and secret proprietary cable ends abound, not to mention Rear I/O ports built into the case. Due to warranty, leasing, and service contract considerations Dell intentionally makes their computers hard to modify.l

 

[uco73]

Hello all. I know this is a bit of a stretch of a question, but i was curious as to whether i could possibly cram a T7910 mobo into my t7600 case. I got a killer deal on the 7600 and have the dual 2687ws with 256gb of ram, but this thing lacks modern cpu instructions amongst other things i have to do to get it to work properly. The dimensions almost look to line up but im not 100% sure. Any help would be appreciated, thank you

In general, Dell computers, like all servers, are not very modifiable. I don't know what problem you had with the T7910, but I know that every modification results in a series of problems.

 

[SiggiTheHun]

Hello all. I know this is a bit of a stretch of a question, but i was curious as to whether i could possibly cram a T7910 mobo into my t7600 case. I got a killer deal on the 7600 and have the dual 2687ws with 256gb of ram, but this thing lacks modern cpu instructions amongst other things i have to do to get it to work properly. The dimensions almost look to line up but im not 100% sure. Any help would be appreciated, thank you

Simple answer is not likely. While they look like they'd fit, the T7600 motherboard and the T7910 motherboard have several differences that make swapping the T7910 mobo into the T7600 case very difficult. It's mostly a connector problem. You can try it, but be careful and don't expect it to work.

e.g.
Precision T7600 (T7610) controller/harness/backplanes only support 3(6) GB/s SAS speed. Different SFF-plugs, too.
But the T7910 backplane(s) harness assembly just fits the T7600/T7610 case.

Both two memory air shrouds are different between T7600 vs T7910 (click).

Front panel T7600 vs T7910 (click).

Dell 0T31JM 1300W PSU seems to work with T7600/7610/7810 and 7910.

CPU coolers are interchangeble - physical and pin-layout.

 

[lexluthermiester]

e.g.
Precision T7600 (T7610) controller/harness/backplanes only support 3(6) GB/s SAS speed. Different SFF-plugs, too.
But the T7910 backplane(s) harness assembly just fits the T7600/T7610 case.

Both two memory air shrouds are different between T7600 vs T7910 (click).

Front panel T7600 vs T7910 (click).

Dell 0T31JM 1300W PSU seems to work with T7600/7610/7810 and 7910.

CPU coolers are interchangeble - physical and pin-layout.

There are other differences as well.

 

[SiggiTheHun]

Re:
DELL Precision 5820 Tower memory configuration

Maybe someone can explain why - according to user manual P82 - among the other arrangements, the respective three outer DIMM slots can be populated with 32 Gb DDR4 modules for 6 x 32 GB = 192 GB total, whilst all other 8/16/64 GB sizes must be configured to 1, 2, 4 or 8 modules, but not six?

 

[lexluthermiester]

Re:
DELL Precision 5820 Tower memory configuration

Maybe someone can explain why - according to user manual P82 - among the other arrangements, the respective three outer DIMM slots can be populated with 32 Gb DDR4 modules for 6 x 32 GB = 192 GB total, whilst all other 8/16/64 GB sizes must be configured to 1, 2, 4 or 8 modules, but not six?

Good question. Not sure.

 

[Retrorockit]

Dell probably isn't going to say anything about this, so I will guess.
That 192GB configuration fills a gap between 128GB (8x16GB) and 256GB (8x 32GB) capacity. There is probably a slight latency hit to run 6 modules on a 4 memory channel CPU.
Dell only shows configurations that they will offer for sale, and support. There may be many other 6 module configurations that work, but don't serve any real purpose. At least as far as Dell is concerned. It may offer a slightly better price point for large quantity buyers.

 

[lexluthermiester]

Dell probably isn't going to say anything about this, so I will guess.
That 192GB configuration fills a gap between 128GB (8x16GB) and 256GB (8x 32GB) capacity. There is probably a slight latency hit to run 6 modules on a 4 memory channel CPU.
Dell only shows configurations that they will offer for sale, and support. There may be many other 6 module configurations that work, but don't serve any real purpose. At least as far as Dell is concerned. It may offer a slightly better price point for large quantity buyers.

That sounds plausible.

 

[Retrorockit]

In an upgrade situation 4x32GB + 4x16GB would produce the same capacity. Dell isn't showing any mixed capacity configurations either.

 

[SiggiTheHun]

Thx, Retrorockit, as lexluthermiester said above: "That sounds plausible".

I expected a general performance penalty with uneven populated 2DPC channels and couldn´t explain this exception other than maybe some NUMA or whatever related integer multiple was hit with (six) 32GB DIMMS.

Dell probably isn't going to say anything about this, so I will guess.
That 192GB configuration fills a gap between 128GB (8x16GB) and 256GB (8x 32GB) capacity. There is probably a slight latency hit to run 6 modules on a 4 memory channel CPU.
Dell only shows configurations that they will offer for sale, and support. There may be many other 6 module configurations that work, but don't serve any real purpose. At least as far as Dell is concerned. It may offer a slightly better price point for large quantity buyers.

In an upgrade situation 4x32GB + 4x16GB would produce the same capacity. Dell isn't showing any mixed capacity configurations either.

 

[uco73]

Re:
DELL Precision 5820 Tower memory configuration

Maybe someone can explain why - according to user manual P82 - among the other arrangements, the respective three outer DIMM slots can be populated with 32 Gb DDR4 modules for 6 x 32 GB = 192 GB total, whilst all other 8/16/64 GB sizes must be configured to 1, 2, 4 or 8 modules, but not six?

You can use 6 memory modules without any problems. Just place them in order from the ends towards the processor, which means 3 above and 3 below the processor. Leave the slots right next to the socket empty and your computer will work without any problems.

 

[DMFCAVU]

New guy on the block, this looks like a friendly bunch so I thought I would drop in with a question whose answer has been eluding me for a while. This is my first foray into water cooling, something I swore I would never do. Until I started processing photogrammetry imagery. A 3300 file dataset took 5+ days of flat out, wide open processing between the XEON W-2295 and dual/NVLINK'ed NVIDIA Quadro RTX 5000's. While it throttled only a couple of times, the CPU temps made me uncomfortable, enter the water cooling idea.

Bought the Koolance EX2-1055 system, somewhat of a hybrid between an AIO and custom system. Its footprint was exactly what was needed. Installation was easy, until I came to the "Slot Adapter". It requires connections or placement of these three items - Temp probes (3) of them. Simple, no wiring, just place the probe where you want to measure - CPU, RAM, GPU. Power - 12 volt, 2 pin Molex connector. They offer an optional external module (which I chose) that addresses the power issue. The termination end of the temp probes now goes to header pins located on the external module, attached to the rear of the EX2. This leaves the 3rd connection - an "ATX Pass-through", whose function is to trigger a "shutdown" command when/if the CPU's (or other measured component) temp exceeds a preset threshold. And therein lies the rub.

This pass-through connection is simple enough - for most systems. Simply connect the "male" portion of the lead to the system's main chassis power button. The remaining "female" connector goes to the motherboard's power switch connection, typically labeled PWRSW, PWSW, or PWBT. As straightforward as this is, it's anything but as far as the 5820 is concerned.

There is a header on the MB labelled PWR_REMOTE, the female connection would go to (logically, but anything but confirmed). The power switch on the 5820 is hard wired into the system via a daughterboard, which then connects to the main board by a ribbon cable.

Long story, but hopefully makes the dilemma clear. Does anyone have any suggestions for tapping into the power switch wiring to complete this circuit? I've purchased a spare daughterboard to experiment on - whether it's performing surgery on the power switch, the ribbon cable, de-soldering, resoldering, etc. All suggestions are welcome.

I've even considered an external PCIe card chassis to provide an option to install a Teradici Remote Access Host card that accommodates the remaining connector to preserve the remote power on/off capability. Just not certain it will work if the Teradici service is not active.

Any suggestions are appreciated.
Thanks,
Dennis

 

[LoveLiveHonoka]

For anyone with the Dell T5810 with WINBOND 25Q128FVSG BIOS chip that needs Turbo Boost Unlock modified BIOS, I made one here. Flash with CH341a programmer. Disable C-States in BIOS after flash.
Flashing process is very easy, you can follow the YouTube tutorial uploaded by Miyconst. I did not need to remove the CPU, I only removed PCIe devices to make the BIOS chip more accessible, and had PSU plugged in. For people with Macronix International MX25L12839F BIOS chip, use Miyconst's BIOS.

 

[SiggiTheHun]

Hello there and welcome.
Branch off the intrusion switch (top connector on the I/O-board or even on the intrusion switch itself? Or is this too much of an emergency-shut off?

https://koolance.com/files/products/manuals/manual_ex2-755,1055_d110eng.pdf P12
https://dl.dell.com/content/manual3...-5820-tower-owner-s-manual.pdf?language=en-us P44-48

 

[lexluthermiester]

New guy on the block, this looks like a friendly bunch so I thought I would drop in with a question whose answer has been eluding me for a while.

Cool! Welcome to TPU!

Bought the Koolance EX2-1055 system

Impressive!

This leaves the 3rd connection - an "ATX Pass-through", whose function is to trigger a "shutdown" command when/if the CPU's (or other measured component) temp exceeds a preset threshold. And therein lies the rub.

This pass-through connection is simple enough - for most systems. Simply connect the "male" portion of the lead to the system's main chassis power button. The remaining "female" connector goes to the motherboard's power switch connection, typically labeled PWRSW, PWSW, or PWBT. As straightforward as this is, it's anything but as far as the 5820 is concerned.

There is a header on the MB labelled PWR_REMOTE, the female connection would go to (logically, but anything but confirmed). The power switch on the 5820 is hard wired into the system via a daughterboard, which then connects to the main board by a ribbon cable.

If I understand you correctly, you want the pump system to come on when needed and shut off when not? I'm not thinking of anything that the 5820 has built in that could seamlessly interface beyond an RS232 port that can be used to interface with the pumping unit, which is one the far end of the unit. A simple PCIe serial card would do the trick. The software is the concern and I couldn't find any in the documentation on the Koolance website fir that unit.

Does the unit have a power-presence actuation function? If so, you could just rig a molex line adapter to run out to the pump unit. In this way the pump would be on when the system is powered, and shut off when it's not.

Beyond that, you may just have to be happy with manually turning both the system and the pump unit on and off at the same time. The pump unit, from what I read, is rated for 24/7 use, so you don't have to worry about wearing it out.

Hello there and welcome.
Branch off the intrusion switch (top connector on the I/O-board or even on the intrusion switch itself? Or is this too much of an emergency-shut off?

https://koolance.com/files/products/manuals/manual_ex2-755,1055_d110eng.pdf P12
https://dl.dell.com/content/manual3...-5820-tower-owner-s-manual.pdf?language=en-us P44-48

I don't think that'll work, but there's no harm in giving it a try.

 

[DMFCAVU]

Hello there and welcome.
Branch off the intrusion switch (top connector on the I/O-board or even on the intrusion switch itself? Or is this too much of an emergency-shut off?

https://koolance.com/files/products/manuals/manual_ex2-755,1055_d110eng.pdf P12
https://dl.dell.com/content/manual3...-5820-tower-owner-s-manual.pdf?language=en-us P44-48

Wow. That is the stroke of genius I've been looking for. That's some serious out of the box thinking, thanks for the suggestion. I'll be trying it tonight, will update with the results. That may actually solve two problems - the first being the intrusion switch no one has been able to figure out how to disable since Dell will not offer any guidance on how to render it inop. I've tried a jumper on the switch terminals, no effect. Changing the BIOS setting had no effect either. Too much of an emergency? I think not. The instant the side panel is opened, it's an immediate shut down.

 

[lexluthermiester]

I'll be trying it tonight, will update with the results.

Yeah, let us know. It's an intriguing idea.

That may actually solve two problems - the first being the intrusion switch no one has been able to figure out how to disable since Dell will not offer any guidance on how to render it inop. I've tried a jumper on the switch terminals, no effect. Changing the BIOS setting had no effect either. Too much of an emergency? I think not. The instant the side panel is opened, it's an immediate shut down.

BTW, simply shorting the connectors together defeats the whole thing.

 

[DMFCAVU]

Cool! Welcome to TPU!

Impressive!

If I understand you correctly, you want the pump system to come on when needed and shut off when not? I'm not thinking of anything that the 5820 has built in that could seamlessly interface beyond an RS232 port that can be used to interface with the pumping unit, which is one the far end of the unit. A simple PCIe serial card would do the trick. The software is the concern and I couldn't find any in the documentation on the Koolance website fir that unit.

Does the unit have a power-presence actuation function? If so, you could just rig a molex line adapter to run out to the pump unit. In this way the pump would be on when the system is powered, and shut off when it's not.

Beyond that, you may just have to be happy with manually turning both the system and the pump unit on and off at the same time. The pump unit, from what I read, is rated for 24/7 use, so you don't have to worry about wearing it out.


I don't think that'll work, but there's no harm in giving it a try.

Cool! Welcome to TPU!

Impressive!

If I understand you correctly, you want the pump system to come on when needed and shut off when not? I'm not thinking of anything that the 5820 has built in that could seamlessly interface beyond an RS232 port that can be used to interface with the pumping unit, which is one the far end of the unit. A simple PCIe serial card would do the trick. The software is the concern and I couldn't find any in the documentation on the Koolance website fir that unit.

Does the unit have a power-presence actuation function? If so, you could just rig a molex line adapter to run out to the pump unit. In this way the pump would be on when the system is powered, and shut off when it's not.

Beyond that, you may just have to be happy with manually turning both the system and the pump unit on and off at the same time. The pump unit, from what I read, is rated for 24/7 use, so you don't have to worry about wearing it out.


I don't think that'll work, but there's no harm in giving it a try.

Really was just focused on the over-temp shutdown. Not concerned about the cooler continuing to run, especially if the system shuts down due to high temps. Kind of like turbochargers on cars/bikes - last thing you want is a motor shutdown immediately after running wide open. The oil circulating through the turbo cools it - an immediate shutdown while hot (red hot) will bake the oil, clogging oil lines and shortening the turbo's life. OK, I know it's a stretch, but in the cooler's case, circulating coolant carries heat away from the CPU until it's at a much lower temp. I don't believe I'll encounter this scenario, but you never know until work the system as hard as it can. An update on success or failure of this approach will be forthcoming! Thanks to all!