"Rate my Rig" thread

[The Infidel]

"IRQ not less than or equal" is usually a driver issue (or anti-virus/anti-malware services), but I have had them pop-up due to hardware failures (a blown capacitor on my board). You said you upgraded your video card, is this about the time you started noticing the blue screens?

Either way, I would try a complete uninstall of the Geforce Experience and Drivers and do a fresh install. After that, look for any updated drivers for your motherboard chipset, NIC, and audio. Also, you could update the driver for your new monitor, but that's a stretch.

There could be corrupted Windows drivers, which would mean you have to mess around with "verifier.exe," but I'll pass on that for now.

The IRQ message is just one of many very different messages with seemingly little in common. I haven't collected the messages yet. I remember another message said something about memory management, but there are other messages as well. These messages have been around since I bought the mobo/cpu/ram, more than 2 years ago. Usually, they are just an annoyance, since it happens when my computer comes back after sleep and after the blue screen reboot, the computer still reads the hibernation data, so I lose nothing. Only a very few times has the blue screen happened during normal operation. If there had been some consistency in the messages, it would give me something to work on. The only thing they seem to have in common, is that they mostly pop up after hibernation.

They annoy me because I can't find the source of these blue screens. I'm using some rather old programs to test the system, from http://www.ultimatebootcd.com/ and a few tests on Aida64. Found this, maybe time to check some other programs? https://www.techpowerup.com/memtest64/

[Ace Pace]

I'm really really against manual diagnosis without any "expert" knowledge, but here goes technical advice over the internet.

Windows 8.1+ BSOD are not common. If they are, you either have
1) Malfunctioning hardware
2) Crappy software

Lets assume you're not running some custom drivers or shit (and running Win x64, meaning no unsigned crappy drivers), it's going to be hardware.

If you can perform a good memtest (anything that covers the full 64-bit address range, Memtest and memtest+ both do), along with furmark (a GPU test) and good CPU test (either Prime95 or compiling something large).

My guess is that those are going to peak something (after an hour+ of runs).

You can also send me some of the crash dumps (MEMORY.DMP under C:\windows) and I can take a look if there's anything in common.

[Starglider]

The new AMD product stack (Ryzen, Threadripper, EPYC) looks great. Unfortunately what I am looking for is moderate core count moderate clocked dual socket. Threadripper is single socket only and so far EPYC is high core count low clocked only. A single Threadripper will be destroyed by a pair of Xeon 6146s. Hopefully AMD will release an EPYC 5-series with Threadripper equivalent core count & clock speeds before I need to buy more workstations. EPYC still great for typical servers, although currently I am buying more beefy quad socket servers mostly because of the horrible distributed shuffle scaling of Spark. Yes dual EPYC has nearly as many cores as a typical prev-gen quad socket, but the bar is about to increase to quad 8168s.

[The Infidel]

I'm really really against manual diagnosis without any "expert" knowledge, but here goes technical advice over the internet.

Windows 8.1+ BSOD are not common. If they are, you either have
1) Malfunctioning hardware
2) Crappy software

Lets assume you're not running some custom drivers or shit (and running Win x64, meaning no unsigned crappy drivers), it's going to be hardware.

If you can perform a good memtest (anything that covers the full 64-bit address range, Memtest and memtest+ both do), along with furmark (a GPU test) and good CPU test (either Prime95 or compiling something large).

My guess is that those are going to peak something (after an hour+ of runs).

You can also send me some of the crash dumps (MEMORY.DMP under C:\windows) and I can take a look if there's anything in common.

I'm grateful for the advice.
I'm running Windows 10 Pro 10.0.14393.1480 64bit and try my best to keep a nice and clean computer. I have this little baby. It is kinda big.

Code: Select all

05-Aug-17  19:02     1,719,033,835 MEMORY.DMP

I'm gonna try some of these programs, it is a while since last time, but I need time for it first, might take a couple of days.
Furmark, is that free? How is that different from 3Dmark or Valley benchmark? (Or hours of gaming. Usually stuff works fine then.)

... Found it... Tried it, but aborted after a few minutes, as my GPU passed 85c. My new 1080 has a stock cooler. My old 680 has a huuuuge custom cooler and never went hot and I still had the problem then. My CPU test has been y-cruncher lately, since it calculates pi using every core available. Every core goes to 100% instantly, but stops after about 15 minutes. Downloaded Prime95 and am running a blended test now. It's using about 2Gb of my memory and stressing cores at about 85%. So far, CPU is about 55c stable.

[The Infidel]

Interesting... A bit into Prime95, my CPU slowly got up to 83c, then it rebooted. My CPU fan maxes out 100% from 70c.

[Mon Aug 07 22:58:20 2017]
Self-test 448K passed!
Self-test 448K passed!
Self-test 448K passed!
Self-test 448K passed!
Self-test 448K passed!
Self-test 448K passed!
Self-test 448K passed!
Self-test 448K passed!
Self-test 8K passed!
Self-test 8K passed!
Self-test 8K passed!
Self-test 8K passed!
Self-test 8K passed!
Self-test 8K passed!
Self-test 8K passed!
[Mon Aug 07 23:03:39 2017]
Self-test 8K passed!
Self-test 512K passed!
Self-test 512K passed!
Self-test 512K passed!
Self-test 512K passed!
Self-test 512K passed!
Self-test 512K passed!
Self-test 512K passed!
Self-test 12K passed!
Self-test 12K passed!
[Mon Aug 07 23:08:50 2017]
Self-test 512K passed!
...reboot...

This, I think, might be something else, as blue screen 99% happens after hibernation and my computer is still not warm. Maybe I need to improve my CPU fan. I have a Thermathake Archon with TT 140mm fan. It usually runs very smoothly and I've never seen temps this high before. I have an exit fan in my cabinet not too far behind. This is a 120mm unit going at low speeds. Maybe time to upgrade to a 140mm with some more omph at 100%?

[Ace Pace]

This looks fine. The reset is very strange. Temperature doesn't sound a likely culprit. We're down to asking generic questions like up to date BIOS, drivers?
If you can upload the dump, I can take a look.

[The Infidel]

This looks fine. The reset is very strange. Temperature doesn't sound a likely culprit. We're down to asking generic questions like up to date BIOS, drivers?
If you can upload the dump, I can take a look.

I appreciate the very generous offer, and you're a long lasting member of this board, but still... I can't conclusively deny that the memory dump won't contain unencrypted passwords that I use, since I at any time usually use at least 10 services that need me to log on, like gmail, microsoft account, facebook, etc., so I respectfully decline your offer.

I might be a bit paranoid about this, but I've always been careful with security, and a year as a security analyst at a SOC didn't really make things better.

[The Infidel]

I use the latest, stable BIOS that was released to this motherboard and try to have my software and drivers up to date. I prefer a clean system and have just recenty installed APP Center from Gigabyte to manage my fans and the very light overclock I tried, then the underclock, but this had no effect on my devious blue screen. A guy at work had some fans he didn't use, so I might try something there, but I doubt it will do much. Currently my mobo is at 30c, cpu 22c(cores around 30c), PCH 38c and GPU 33c. All my disks hover around 30c. All temps reported by Aida64.

[Ace Pace]

If you have unencrypted passwords in memory, switch browsers.

I use the latest, stable BIOS that was released to this motherboard and try to have my software and drivers up to date. I prefer a clean system and have just recenty installed APP Center from Gigabyte to manage my fans and the very light overclock I tried, then the underclock, but this had no effect on my devious blue screen. A guy at work had some fans he didn't use, so I might try something there, but I doubt it will do much. Currently my mobo is at 30c, cpu 22c(cores around 30c), PCH 38c and GPU 33c. All my disks hover around 30c. All temps reported by Aida64.

It's not temperature and it doesn't sound like software.

BIOS settings, did you tweak anything? Not CPU OC settings, but memory settings (default, XMP or did you play?). Is it configured to normal boot, fast boot or ultra fast? Dynamic Storage Accelerator on/off? I'm looking at your motherboards manual now and

I'm looking at what you wrote about hibernation and I'm trying to think of what combo of features did you play with that maybe screwed it up. If you have a separate SATA controller (I think this motherboard has one), try moving your hard disks over there and see if that helps.

[Anacronian]

To me it sounds like you are getting a reset due to the CPU getting too little power, It can happen if you are too conservative with your power setting.

I suffered the same problem for a while and the solution was to bump up my power a few volts.

[The Infidel]

OK, something "interesting" happened while checking memtest86, but using cpu's sequential instead of all at once (default). Memtest86 froze. Tried round robin twice, and it froze at the same time and the same pattern. Tried normal sequential twice, and it froze at a different time than round robin (same time for sequential runs), but same pattern. This was not during the hammer test, so it would not thaw and I had to reboot. I don't think it likes DWDWDWDW... The address test (own address), is that when it checks





Also, a new blue screen when waking up from sleep: "Memory management". I think I'm onto something now. Time to check my memory modules one by one. It's not that much of a irritation moment, since it just makes the waking up take about 15 seconds longer, but I just have this urge to find out now.

[The Infidel]

To me it sounds like you are getting a reset due to the CPU getting too little power, It can happen if you are too conservative with your power setting.

I suffered the same problem for a while and the solution was to bump up my power a few volts.

Well, everything is running stock now. No overclock. This system has only been overclocked about 30 minutes.

[The Infidel]

If you have unencrypted passwords in memory, switch browsers.

I use the latest, stable BIOS that was released to this motherboard and try to have my software and drivers up to date. I prefer a clean system and have just recenty installed APP Center from Gigabyte to manage my fans and the very light overclock I tried, then the underclock, but this had no effect on my devious blue screen. A guy at work had some fans he didn't use, so I might try something there, but I doubt it will do much. Currently my mobo is at 30c, cpu 22c(cores around 30c), PCH 38c and GPU 33c. All my disks hover around 30c. All temps reported by Aida64.

It's not temperature and it doesn't sound like software.

BIOS settings, did you tweak anything? Not CPU OC settings, but memory settings (default, XMP or did you play?). Is it configured to normal boot, fast boot or ultra fast? Dynamic Storage Accelerator on/off? I'm looking at your motherboards manual now and

I'm looking at what you wrote about hibernation and I'm trying to think of what combo of features did you play with that maybe screwed it up. If you have a separate SATA controller (I think this motherboard has one), try moving your hard disks over there and see if that helps.

I haven't tweaked the BIOS, but now I'm starting to wonder if the guy I bought it from did something. I don't remember if I resat the BIOS or not. Don't think I did and my system is doing just fine 99% of the time. I did recently change from normal boot to fast boot, but have turned back to normal boot again. DSA? Have to check that one.

[TheFeniX]

Got bored last night and looked around for how much a second Xeon would cost for my aging Dell T610. Found a refurb for $5 shipped... ok. And a stock cooler (just a giant radiator, airflow is controlled by two large and extremely loud back fans) for $20 shipped. Normally, I would have dug around, but I know this one will fit an getting a cooler for $20 without having to drive doesn't seem that annoying. Something that gave me a chuckle was the fan noise on the 610s has been a complaint for years. Just recently, like in July this year, Dell released an iDRAC (whole system is controlled by it and mostly unconfigurable) firmware update that increases the threshold to bump the fans up to "afterburner."

Anyways, if the CPU is garbage, I'm out 25 bucks.

Not that I really need it, but should be fun to mess with.

Now if I could only just find another 16GBs of ECC DDR3 RAM for under $160. But, either way, I'm going to have to find that original 4GBs of ECC RAM I pulled out.... or dig something else up. There's some refurbed 8GB ECC sticks I've found for $35 a piece, but they're 1.5V vs the 1.35V I've got (which is much more expensive) and I don't know if I want to mess with under/over voltage. Plus, it's slower. Oh well, I'll work something out.

Really just mucking about, but I plan to look into a local DNS adblocking solution and figured Linux was the way to go, so I'll just VM a distro. Was thinking Debian since it's what that Rhasberry PI thing Infidel was talking about is based on.

[The Infidel]

That DNS adblock is very easy to love, once you tried it.

OK, have tried memtest on all modules, two modules, single module put into different slots and channels. All the same: No errors reported running all cores (default) or if I choose a core, no matter which one I try. If I try to run round robin or sequential, memtest freezes no matter memory module(s) tested or configuration. It might be a firmware thingy, but I am running the latest BIOS that came (in 2015).

You know what? Fuckit (until futher notice). I've used this computer more than two years in it current state. Bluescreen happens almost only after hibernation and only about 10% of the time, Prime95 is the only thing that makes my CPU boil. Furmark is the only thing that makes my GFX boil. I've changed one of my case exhaust fans from an old PABST 120 to a Fractal Design 140.

[TheFeniX]

So I vented about this in the thread, but someone might find it useful.

The W10 Fall Creators update makes a shitton of changes. I jumped in early for the fixes to DX9 games once again (as this was working in W7, but not W8 or W10) give you access to pretty much all your VRAM and can even pull from system memory for rendering assets. This might cause some performance issues, but if you run gigabytes of high-res texture and high-poly models for a game like Skyrim, at the least you won't crash.

And let's face it, any moderately powerful GPU can bang out like 200FPS on Skyrim anyways.

Well, after all the setup, I was pulling nothing but 30FPS in a lot of games. Heroes of the Storm would pull 140FPS, but that's likely due to in running borderless windowed. Any full screen game was capped at 30FPS and the weird part was that it WOULD go up to 60FPS for a few seconds, then slowly just work back down to 30FPS. I had already disabled the XBOX Game DVR and GameMode options. Google was no help.

Well, it turns out having an old (like 2007, at the latest) Sony Bravia 1080p TV hooked up at all, even turned off, was causing the issue. Which is weird because I'm pretty damn sure the thing is 120hz (60, doubled). But whatever. So, if you find yourself with your framerate locked for no fucking reason, check your connected HDMI devices. I would assume this only affects HDMI and I also assume it's some HDCP bullshit.

[Dominus Atheos]

So I vented about this in the thread, but someone might find it useful.

The W10 Fall Creators update makes a shitton of changes. I jumped in early for the fixes to DX9 games once again (as this was working in W7, but not W8 or W10) give you access to pretty much all your VRAM and can even pull from system memory for rendering assets. This might cause some performance issues, but if you run gigabytes of high-res texture and high-poly models for a game like Skyrim, at the least you won't crash.

And let's face it, any moderately powerful GPU can bang out like 200FPS on Skyrim anyways.

Well, after all the setup, I was pulling nothing but 30FPS in a lot of games. Heroes of the Storm would pull 140FPS, but that's likely due to in running borderless windowed. Any full screen game was capped at 30FPS and the weird part was that it WOULD go up to 60FPS for a few seconds, then slowly just work back down to 30FPS. I had already disabled the XBOX Game DVR and GameMode options. Google was no help.

Well, it turns out having an old (like 2007, at the latest) Sony Bravia 1080p TV hooked up at all, even turned off, was causing the issue. Which is weird because I'm pretty damn sure the thing is 120hz (60, doubled). But whatever. So, if you find yourself with your framerate locked for no fucking reason, check your connected HDMI devices. I would assume this only affects HDMI and I also assume it's some HDCP bullshit.

Windows 10 new "game mode" forces v-sync, locking your fps to 30 or 60 to prevent screen tearing. Turn it off by going to settings >Gaming >Game Mode. You can also toggle it in-game by hitting the Windows key + G.

[TheFeniX]

Yea, that and the XBox DVR shit was the first to go way back during the first Service Pack (or whatever the Hell they call it now). What killed me is every search result was like "disable THESE!" and I even went as far as registry edits. I was even force closing RivaTuner and the GeForce recording and playback services manually to no avail. At one point, pretty much anything W10 would let me close out was closed out.

That's why it was just so damn weird that (even if old) disabling a 120hz TV is what fixed it. I can only assume the dated tech from my TV makes W10 force my GPU into a compatibility mode and lock it to 30FPS. Either way, I'm not risking plugging the damn thing back in.

[LadyTevar]

WV Broadband Speed Test results
Someone with more knowledge tell me how this rates?

[TheFeniX]

Your question is relative based on location and cost. In general, I've seen worse, but also much much better. I'm paying out the ass like $100 a month, but I'm getting 80Mb/s Down and 10mb/s up. Since both my wife and I work/play online a lot and only stream our entertainment, it's "worth it" in that even cutting it down to $60 a month package just eviscerates what bandwidth they would give us. Divide by 8 to get MB and we're pulling just at 10 meg a second, which is nice.. until you talk to people in other countries who pay like $20 a month for Fiber.

I just downloaded Dark Arisen (again) in under 30 minutes. The game is 22GBs.

Meanwhile, my brother out in Sealy pays $100 a month for 20Mb/s directed wireless and, through me helping him test, found out they are throttling the shit out of his bandwidth after the first couple megabytes. All the way down to 5Mb/s. This is a shitty way Internet companies beat "speed test" websites.

In my area, Comcast offers a 10Mb/s package for about $30. As said, prices and speed (and how much they screw you) varies by location.

[Jub]

For a comparison, I'm in Vancouver getting these results for $49.90 year one, $84.90 year two, and $105.00 as a regular rate once the promo runs out.

http://broadbandwv.speedtestcustom.com/ ... 0467fe9386

http://www.speedtest.net/my-result/6740672560

For reference, we're paying for a 150 down 15 up package and getting these results.

[Zaune]

PCPartPicker part list / Price breakdown by merchant

CPU: Intel - Core i5-7600 3.5GHz Quad-Core Processor (£183.54 @ Aria PC)
Motherboard: Asus - STRIX Z270-E GAMING ATX LGA1151 Motherboard (£177.50 @ Aria PC)
Memory: Kingston - HyperX Fury Black 16GB (2 x 8GB) DDR4-2133 Memory (£129.99 @ Ebuyer)
Storage: Western Digital - Green 240GB M.2-2280 Solid State Drive (£71.39 @ Aria PC)
Storage: Seagate - Barracuda 3TB 3.5" 7200RPM Internal Hard Drive (£81.39 @ Eclipse Computers)
Video Card: PowerColor - Radeon RX 570 4GB Red Devil Video Card (£191.36 @ Ebuyer)
Case: Phanteks - Enthoo Pro ATX Full Tower Case (£94.25 @ Amazon UK)
Power Supply: SeaSonic - 450W 80+ Gold Certified Semi-Modular ATX Power Supply (£62.99 @ Overclockers.co.uk)
Optical Drive: Lite-On - iHAS124-14 DVD/CD Writer (£10.50 @ Aria PC)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Case Fan: ARCTIC - Arctic F12 Pro 54.0 CFM 120mm Fan (£4.99 @ Amazon UK)
Other: 40cm White LED Strips x2 (£9.99)
Other: Phantex PWM Fan Hub (£19.32)
Total: £1067.15
Prices include shipping, taxes, and discounts when available
Generated by PCPartPicker 2017-12-09 23:48 GMT+0000

This is hopefully going to be what I spend my Christmas money on.

[Dominus Atheos]

That's a very high-end a motherboard to pair with that very mid range processor. Save money by going for an upper midrange mobo and then bump the proc up to an upper midrange unlocked 6-core coffee lake.

https://www.aria.co.uk/SuperSpecials/Ot ... ctId=68632

https://www.aria.co.uk/SuperSpecials/Ot ... ctId=68439

Although you are clearly going for an e-peen build, with £55 spent on case cooling.

[Zaune]

What can I say? The case has all those fan hardpoints, so I might as well use them. And if it was that much of an "e-peen" build I'd buy a ridiculous amount of ugly multi-coloured LED lighting gadgetry as well. I just want to spend a bit more money up front and not need to upgrade for another five years or buy a new rig for ten.

Anyway, I couldn't find that exact motherboard on PCPartPicker but I did find this reasonable-looking one for under a hundred. Combined with an i5-8400 (because I'm not yet brave enough to try overclocking) I think that gets me slightly better performance and brings the price tag under the thousand pound mark.

[TheFeniX]

I'd also have to recommend at least the 6-core as even though PC ports have come a long way, that's what the current consoles have.... at like 1Ghz or something stupidly low. Though, as a joke I threw my 970 into the dual-core setup I built for my nephews and even that thing could pull 40FPS in Ark, with turning down some of the more CPU intensive aspects. To be fair, it was OCed to like 4.2GHz. They had to live with the 660Ti though.

You run Linux right? But this is also going to be a gaming rig? That 240GBs on the SSD is going to go fast. Any game you can put on an SSD is money. Any game on the HDD will feel like you're booting into Windows 2000 or 95 comparatively. Or playing Half-life/Morrowind. LOADING.... LOADING..... LOADING.... I pushed Fallout 4 back on my HDD to save SSD space, not that the game was worth a shit before that, but I can't even be bothered to wait for the load times. I just don't play it.

In lieu of that, you might dump money into a second HDD, or cut down to 2 1TB drives and stripe them. Haven't messed with striping since BF1942 was a thing, but the straight read performance benefit was immediately apparent.

EDIT: striping SSDs is....ok. I had two cheap 120GBs I threw into my old rig just so I could put windows and other junk into one drive. The extra speed is there, but not enough to make you dance a jig.

And man, holy shit I can't wait to get another few SSDs. Trying to finish out my degree: Virtual Machines on an SSD are in a whole other league. My god, server's with multiple services starting in under 60 seconds. Updating one, installing services on another, it just goes. Meanwhile, my HDD is dieing just trying to load Windows 7.