-Ice

Thanks for the tip on quoting! As for the HDD light, I have a corsair carbide 500R case and it has a power button that lights up and below that is a HDD activity light. I guess my question is what drives the power button light if the power button wire just transmits an on/off signal?

Thanks! No need for LED really, I just want a push button switch. Is the HDD activity light different from the power on/off light?

Bump! Any suggestions on wire gauges and switch?

Yep, that's it!

That is insane! I didn't even know these things were available! I'll have a look into this to see if I like it but knowing me, my next issue would be losing that controller! Thank you for the offer of the diagram, please do if nothing but to confirm what I am thinking of. I'm no expert in wiring or electronics so I need all the help I can get! EDIT: after reading up a bit, it seems that it can have some issues with setting up and connectivity so my hardwire idea would still be the most reliable.

I've never done this before so really want to be careful with this. I'd like to make a secondary/alternative power switch for my PC. Kind of like this but instead of a trick switch, it'll just be a second power button: https://www.youtube.com/watch?v=yBj02EBj_dQ I've re-designed my simpit and my case is a bit further away than I'd like so I was thinking of putting a power switch inside the simpit that's more accessible. I'm thinking just a momentary switch (pushbutton), run the wires through and behind the simpit and to the PC. However, looking at wires and momentary switches on Amazon and I quickly realize there may be more to this than I've expected. Do I need a particular gauge or size of wire? Do I need a specific momentary switch (pushbutton) for this application? I am also thinking of a connector between the switch (and its wiring) and the PC itself. The idea is to connect/solder this connector to the PC case's wiring for a permanent connection and then this connector exits out the back of the case where I can easily connect/disconnect the switch as needed, ie for maintenance purposes. Any ideas for types of connectors I can use? I'm based in the UK so if anyone can point me to local stores you've bought from or trust, or maybe Amazon, that'll be great! Any help appreciated! Thanks!

Here's a screenshot of the testing I did before: And here's one I ran just now. Temp spiked to 53 Celcius. Why is the 4K write so low and the 4K Q32T1s read and write so different from before? However, is what does that have anything to do with the laggy behavior I see on the desktop?

Copying a video file was mostly around 60MB/s but peaked around 150MB/s around the start. Copying within the NVMe was fast, around 300-600MB/s until the end where it went to around 9MB/s but still fast enough to be noticeable compared to HDD-to-NVMe copy. Within HDD was fast as well. I fail to see how it can be a GPU issue. I'm on a 980Ti with latest drivers. But I'll reinstall drivers later to test anyway. As for caching confusion, well, like I said, I used to have the same config with a SATA SSD instead of NVMe drive with no issues. Was also running some tests on the NVMe drive earlier using CrystalDisk Mark and the NVMe drive went up to 51 Celcius... probably went up higher than that as well, it only registered 51 as CrystalDisk Info refreshed at that point. I've changed it now so that it refreshes every minute. The drive runs hot... 38C on idle whereas my SSDs are 23C (850 Evo). My HDDs are at 25-26C. My external drives are 32C and 36C. So for a "stick of gum" that's over my motherboard that gets good airflow, idling at 38-39??

Finally finished installing Win10 Anniversary edition on my Samsung 960 Evo 250GB NVMe drive. Boot up is faster by a few seconds, but I'm having a bitch of a time arranging my desktop icons. I have Fences installed but even on icons that are not in Fences, everything is just laggy. Where before I could just click on an icon then drag it to where I want it, now I have to click the icon, make sure there's a box around it, move my cursor to where I want it, let go of the cursor, then watch my icon move to that location, then watch as the other icons re-arrange themselves. Dragging Fences around is also laggy as hell; I don't think I had this issue even when my OS was on a HDD. My OS is on the NMVe drive, my deskto "folder" is on a HDD, but is the same location I used when my OS was on my old Crucial M4. So why would it be laggy on the NVMe but not on the M4 considering the "folder" is located in the same place?

Wow... 54,708 hours of power on time? That's 6.24 years of being ON, and only switched on 408 times? I'm curious what that drive is for? NAS?

If you wanted more speed, why stick to SATA? Why not PCIe NVMe drives?

Back in August, my Samsung 840 just decided to quit. I was upgrading to Win10 and on one boot up, everything was working fine, the next restart and the drive just disappeared. Not detected in BIOS, not detected in Disk Management. DEAD. It was just over 3 years old and was used as my Games install drive. I only installed my favorite games on it, so it didn't get very much data written on it and I was surprised for it to die at that time. Yesterday, my Crucial M4 decided to quit as well. I was installing a case fan on my PC and on reboot, the system told me it could not find a boot drive. Great. The drive could not be detected in BIOS. I plugged it in to my son's computer and confirmed that it was not seen in BIOS and in Disk Management. DEAD. The annoying bit was that this was my OS SSD. I initially panicked as I kept a few documents on my Desktop for easy access, then I remembered that I "relocated" my user files (Desktop, Documents, Videos, etc.) to my HDDs, but was still quite annoyed that stuff like these would be working 100% fine one moment then just die the next boot up. I purchased this drive on April 2012, so 5 years of good service was a good run, but still didn't expect it just to go like that. Is this supposed to be normal SSD behavior? As the warranty was obviously out, I decided to open up the SSD just to have a look inside. Nothing looked burnt or worn out, so I re-assembled the drive. For some reason, I decided to plug it in again on my son's computer, not expecting anything, and to my great surprise, the PC started booting up in Win10!! My son's PC has been rolled back to Win7 so I immediately knew it was my Crucial drive that was working. It's ALIVE!!!! I plug it back into my PC and it starts back up like nothing's happened! 4 freaking hours of troubleshooting, trying different SATA cables, different power cables, and it just boots up like nothing's happened!! I noticed my D: and E: drives were swapped around, but aside from this, nothing seems to have changed! I ran Samsung Magician and discovered the new interface has less info than before... but it didn't detect any errors. SMART check via DOS prompt also declares all drives OK. I ran Crystal Disk Info and here's the result: So the SSD should have YEARS of life left in it, if I've only got it down to 97% over 5 years of use as the OS drive. 8,486 hours of power on is 353.58 days, so less than a year of being on. Anybody got a clue what's happening here? I'm reluctant to trust this SSD any more. My plan was to "retire" it and put it in my son's PC as his OS drive (i5 750 setup in sig below); I guess I'll be doing this sooner (next week) rather than later.

What is a "large sequential file copy"??

I'm hoping to "retire" my old Crucial M4 and was thinking of getting either another SSD or an M.2 NVMe for an OS drive. I currently also have a Samsung 850 250GB for games and have short-stroked two 1TB HDDs for other games and use the rest of the drive for storage. I do a little photo editing and would like to do some videos later on once I learn how to do it, but aside from Photoshop and the occasional MS Office work, I usually use my PC for gaming. I've got a strong hunch that most people here would not recommend getting an M.2 NVMe drive since SATA SSDs aren't really being pushed to the limit that often, so what programs or applications would justify buying an M.2 NVMe drive? I'm guessing for boot times and program startups, an M.2 NVMe would not make much difference compared to a SATA SSD, correct? Same thing for game loading times?

Ran another batch of tests, this time with the two touchscreens connected to the 660 Ti. First run, PhysX to 980 Ti - Avg 41.66fps, 660 Ti max Core Load 15%, avg 8.5% Second run, PhysX to 660 Ti - Avg 41.265fps, 660 Ti max Core Load 20%, avg 9.4% Third run, PhysX to 660 Ti Dedicated - Avg 41.67fps, 660 Ti max Core Load 20%, avg 9.8% I immediately removed the 660 Ti which reset PhysX to the only available card, the 980 Ti, and ran a test - Avg 45.863fps I think this effectively shows that the PhysX settings is not what's throttling the FPS numbers. Back to the starting point on this one. For these tests, I've also ran the logging function of HWiNFO so I have that data, just don't know what exactly I should be looking at. As mentioned before, I would take screenshots of the HWiNFO screen before and after the tests so I have that data as well. While it doesn't show real-time GPU utilization, it does show current/min/max/avg numbers.

Okay, ran three tests again... First setup was the 980 Ti and the 660 Ti with nothing connected to the 660 Ti and PhysX set to Auto (it chose the 660 Ti). Avg 41.54fps Second setup was the 980 Ti and the 660 Ti with nothing connected to the 660 Ti and Physx set to 660 Ti and ticked the "Dedicated to PhysX" box. Avg 41.36fps Third setup was the 980 Ti and the 660 Ti with nothing connected to the 660 Ti and Physx set to 980 Ti. Avg 41.36fps Seems like PhysX has nothing to do with it... Interestingly, HWiNFO reports that the 660 Ti was not under load at all during the last test!

Hmmm.... could be. I wonder if I plugged the card back in... but forced it to use the 980 Ti for PhysX? Back to the drawing board!! The original came out in 1998. BMS is a mod that does it's darnest best to bring it up to more modern standards. How do I monitor the utilization of each GPU? As for the magic bullet, no worries. I really appreciate your input and you've just given me one more thing to test. Hopefully, I'll find some sense in this... I don't care if at the end of the day I can't use the 660 Ti, I just want to find a good explanation WHY.

As you can see from the test above, I've had the touchscreens on the 660 Ti and on the motherboard. The performance on the 980 Ti + 660 Ti was the lowest, then 980 Ti + Targus/Motherboard was better by 1-2 frames, but performance on just the 980 Ti was the best of all three.

Okay, done a few tests and trying to make sense of the results. 2 GPUs, 3 screens plugged into 980 Ti, no Helios, no YAME - min 44, max 56, avg 50 2 GPUs, 5 screens plugged in (3 to 980 Ti, 2 to 660 Ti), no Helios, no YAME - min 44, max 56, avg 50 2 GPUs, 5 screens plugged in (3 to 980 Ti, 1 to Targus, 1 to motherboard), no Helios, no YAME - min 45, max 56, avg 50 From these tests, we can see that the addition of the two touchscreens does not really affect the FPS numbers regardless of where the touchscreens are plugged in. During testing, the touchscreens simply showed my desktop wallpaper, so there was minimal load. 2 GPUs, 5 screens plugged in (3 to 980 Ti, 2 to 660 Ti), Helios, YAME - min 36, max 45, avg 41 2 GPUs, 5 screens plugged in (3 to 980 Ti, 1 to Targus, 1 to motherboard), Helios, YAME - min 38, max 47, avg 42 Again, we can see that with the added load of Helios and YAME, the performance is the same regardless of where the touchscreens are plugged in. There is an average drop of 8fps though, due to the additional workload. 1 GPU (980 Ti only), 5 screens plugged in (3 to 980 Ti, 1 to Targus, 1 to motherboard), Helios, YAME - min 40, max 50, avg 45 This is where things get blown out of the water. Minimum FPS is higher, max FPS is higher, average FPS is higher. Sure, there's still a drop in performance due to the added load of Helios and YAME, but **REMOVING** the 660 Ti improves performance? What?!?!?? Here's a graph of the average FPS numbers over the tests: On the first graph, the green line was when the touchscreens were plugged in via Targus and direct to the motherboard. It's interesting to note that the green line is most of the time mid-way between the blue and red lines. On the second graph, the green line is for the test with the 980 Ti only. It is blatantly obvious that the two-GPU setup was nowhere near the single-GPU performance numbers! Also note that the FPS numbers were WORSE when the touchscreens were plugged into the 660 Ti (blue line)!! Why was the 980 Ti hampered by the presence of the 660 Ti? Why were the framerates better when the touchscreens were plugged into the USB/motherboard compared to when they were plugged into the 660 Ti? While this puts the final nail in the coffin regarding the decision to use one or two GPUs, I cannot help but wonder why the 980 Ti was hampered by the presence of the 660 Ti. Remember that on both with and without Helios/YAME tests, I've tested with the screens plugged into the 660 Ti and with the screens plugged into USB (Targus) and motherboard but it was only when I removed the 660 Ti from the PC itself that the FPS numbers went up by a good bit.

The primary monitors are used to display the game which is Falcon 4 BMS and AFAIK, it's on DX9. The two touchscreens display data from a program called Helios for switch control and YAME for gauge display. Helios and YAME do refer to a "shared output" from Falcon 4 BMS in order to do their respective functions, but I expect this to be a CPU-side function and not something that GPUs would deal with. As for "distributing" the processing between the two.... why would the 980 Ti care how fast/slow the 660 Ti can push pixels? Why would it be affected? The two aren't SLI'ed or connected in any way aside from being on the same motherboard....

That is indeed a bit too rich for what I'm planning to do The two touchscreens are only really displaying cockpit gauges and switches, so IMHO does not justify the cost of that splitter. Well, the two cards are nVidia so won't they operate off of the same driver? As for "software overhead," isn't that on the CPU-side of things? With regards to the load, I'm getting 52fps average with just the main displays working (5896x1080 res on three monitors) and I'm getting 45-46fps average with the main displays working plus the two touchscreens (2x 1920x1080 res) so the additional touchscreens do exert a bit more load even though they're not really displaying anything complex. This is even with one touchscreen connected via USB and the other is connected via the motherboard's DP output.

PC specs are as above. Monitors are all 1920x1080. As mentioned above, my three primary displays are connected to the 980 Ti via DVI-to-DP adapter cables. The fourth slot is a HDMI that I use for the HTC Vive, so I still need something to push pixels on the two touchscreens.

I currently have a i5 6600K on a Gigabyte GA-Z170X Gaming 7 motherboard and an EVGA GTX 980Ti. I have three primary screens on Surround setup and two secondary touchscreens. My primary screens connect via DVI-to-DP adapter cables and go into the 980Ti. I used to connect one touchscreen to the HDMI port of the 980Ti, then use a Targus docking station to basically connect my second touchscreen to the PC via USB. I got a HTC Vive last Christmas and so I had to remove the touchscreen from the HDMI port... and got a DVI-to-DP adapter and connected it to my motherboard's DP port. So now, my connections are: three screens to 980Ti, one touchscreen to (Targus) USB, one touchscreen to motherboard, Vive to HDMI on 980Ti. I planned to get a secondary GPU to push pixels onto the secondary touchscreens and get rid of the Targus dock and DVI-to-DP adapter. I managed to get a GTX 660Ti and have connected my two touchscreens to it via DVI cables. I initially connected the 660 Ti to the 4x slot on my motherboard and promplty got issues with my displays. Right-clicking took a while before the menu comes up. Surfing the internet and Firefox became slow. Sometimes, if I leave the monitors on or if I try to changes some settings, my primary screens go dark (loses picture? I can't see my desktop wallpaper or icons!) and sometimes one of my touchscreens go dark as well and the only thing I can do is reboot. The touchscreens would sometimes loose the "touch" feature. I would also get a blue screen and the PC would reboot. I then moved the 660 Ti to the 8x slot and so far everything seems stable although one of the touchscreens seems to like to "reset" by turning off then on again when my main game comes up on the primary displays. I'm running the latest nVidia drivers, PhysX is set to Auto (and uses the 660Ti). Here's where it gets interesting. The main purpose of this exercise was to see if I can get better framerates when playing Falcon 4 BMS. I had run some tests on the initial monitor setup, with only the main displays working, the average was 52fps with max of 58-60fps. With all 5 displays working, the average was 45-46fps and max of 52-54fps. With the new monitor setup and the 660 Ti, all 5 displays on, the average is 41fps and max of 45-46fps. Why is this happening? I was expecting to have better fps results... I was even hoping to get back to the 52fps average... and might even be content if the fps results were unchanged, but why am I getting WORSE results with the 2-GPU setup? Thanks in advance for any advice or suggestions!