The PC Builders Thread ("I Need a New PC") v2

[gdt]

Just finished building my (mostly) new PC. Took about 6.5 hours. No real frustrations. Had a little ram issue at the end, and a mystery cable I figured out was a front panel USB C connector my Mobo doesn't support, but pretty sure I'm all good now. Forgot to download windows onto a USB but doing so now. When that's done I'll boot to that, install that onto the new SSD, and we gon eat.

 

[ploonkey]

I've sent this build earlier this week, but do you think it's a legit upgrade over what I have now?

What I want to build/upgrade to: https://pcpartpicker.com/list/gM2CCL

What I have now:

CPU: i7-6700K Stock
CPU Cooler: Corsair A50 (yes it's that damn old)
MB: Asus Maximus VIII Hero Z170
RAM: 32GB Corsair DDR4-3200 (4x8GB) (Replacing these to go to 3600 so I can do 1:1 with Infinity Fabric)
Storage 1: 1TB Samsung 860 EVO
Storage 2: 4TB HGST 7200RPM (Keeping in the new build)
Video Card: EVGA RTX 2080 (Keeping in the new build)
Case: Phanteks Evolv ATX
PSU: NZXT Hale90 750W 80+ Gold (this thing is over 10 years old)

Click to shrink...

I think you'll only see an upgrade in game performance if your current usage includes streaming.

 

[Dave.]

Question: Can you RMA memory if it's unable to run stable at XMP speeds? I keep getting BSODs at random times exactly like my old computer. Bought a USB stick and tried a memtest and I got 2 errors on test 6. Nothing else is overclocked just the ram at it's xmp profile.

DDR4 just sucks. :/

Click to shrink...

Absolutely yes you can, and should.

 

[Dave.]

I've sent this build earlier this week, but do you think it's a legit upgrade over what I have now?

What I want to build/upgrade to: https://pcpartpicker.com/list/gM2CCL

What I have now:

CPU: i7-6700K Stock
CPU Cooler: Corsair A50 (yes it's that damn old)
MB: Asus Maximus VIII Hero Z170
RAM: 32GB Corsair DDR4-3200 (4x8GB) (Replacing these to go to 3600 so I can do 1:1 with Infinity Fabric)
Storage 1: 1TB Samsung 860 EVO
Storage 2: 4TB HGST 7200RPM (Keeping in the new build)
Video Card: EVGA RTX 2080 (Keeping in the new build)
Case: Phanteks Evolv ATX
PSU: NZXT Hale90 750W 80+ Gold (this thing is over 10 years old)

Click to shrink...

Anandtech bench is a useful tool for getting an idea of how these will compare - there's a number of games at various settings towards the bottom of the list:

CPU 2019 Benchmarks - Compare Products on AnandTech

CPU 2019 benchmarks: Compare two products side-by-side or see a cascading list of product ratings along with our annotations.

www.anandtech.com

 

[Deleted member 9290]

i want to build a "high end as it can be" ITX PC this year.

im now curious what RAM should i use with these parts:

-AMD Ryzen 7 3700X 8x 3.60GHz
- Gigabyte Aorus Pro AMD B450 So.AM4 Dual Channel DDR4 Mini-ITX Retail

im limited to only a handful mainboards because of ITX.
i would just get these ones because they are cheap:

- 32GB G.Skill Aegis DDR4-3200 DIMM CL16 Dual Kit

but i have absolutely now idea about clock latency and whatnot you have to consider for RAM.

can anybody help me?

thanks!

 

[Crazymoogle]

i want to build a "high end as it can be" ITX PC this year.

im now curious what RAM should i use with these parts:

-AMD Ryzen 7 3700X 8x 3.60GHz
- Gigabyte Aorus Pro AMD B450 So.AM4 Dual Channel DDR4 Mini-ITX Retail

im limited to only a handful mainboards because of ITX.
i would just get these ones because they are cheap:

- 32GB G.Skill Aegis DDR4-3200 DIMM CL16 Dual Kit

but i have absolutely now idea about clock latency and whatnot you have to consider for RAM.

can anybody help me?

thanks!

Click to shrink...

Depends what you mean by "high end as it can be" - it's totally possible to go X570 and use higher end processors (ie: 3900X) with mITX builds. Specifically, ASUS and AORUS seem to have good X570 solutions. But since you're looking to get cheap memory, I guess it's not what I think.

That G.Skill kit is totally fine. Ideally, you want DDR-3600 CL16, but that can be quite expensive. DDR-3600 CL18 is often close to price to 3200c16, so it's popular. But 3200cl16 is the most popular and is a totally fine option for Ryzen builds. You should check the motherboard QVL and RAM vendor QVL to see if the two are comparable (if one says it is, good enough) but even if it's not, generally you can try manual timings and make it work. The BIOS will just set your RAM to 2133 by default until you figure it out.

 

[パンダさん]

i want to build a "high end as it can be" ITX PC this year.

im now curious what RAM should i use with these parts:

-AMD Ryzen 7 3700X 8x 3.60GHz
- Gigabyte Aorus Pro AMD B450 So.AM4 Dual Channel DDR4 Mini-ITX Retail

im limited to only a handful mainboards because of ITX.
i would just get these ones because they are cheap:

- 32GB G.Skill Aegis DDR4-3200 DIMM CL16 Dual Kit

but i have absolutely now idea about clock latency and whatnot you have to consider for RAM.

can anybody help me?

thanks!

Click to shrink...

I've bought this exact Mainboard for my GFs PC.
The PC is currently on a 2600x with the 16GB G.Skill Aegis DDR4-3200 DIMM CL16 Dual Kit.
The RAM and the MB are pretty good, the only thing I didn't like, it has only one Fan Header aside from the CPU one.

 

[AnimeJesus]

Just wanted to get some thoughts on the new build I've already ordered:

Current Build (8+ years old) where my CPU is a serious bottle neck:

CPU - i7-870
MB - EVGA P55
RAM - 8GB
GPU - GTX 970
PSU - 500w

New Build:

CPU - i7-8700
MB - MSI Z390-A Pro
RAM - 16GB (Corsair Vengence LPX)
GPU - GTX 970 (will upgrade to a 1060 6GB a little later for free and then a 2060\2080 once they come down in price)
PSU - 650w

Not sure how the new build will compare to the new consoles once they are out and Im also still fine with 1080p gaming as long as my framerates are at least 60. Is this new building going to be outdated in a years time?

 

[Deleted member 9290]

I've bought this exact Mainboard for my GFs PC.
The PC is currently on a 2600x with the 16GB G.Skill Aegis DDR4-3200 DIMM CL16 Dual Kit.
The RAM and the MB are pretty good, the only thing I didn't like, it has only one Fan Header aside from the CPU one.

Click to shrink...

Cant you just buy another Y-cable for more fans?

 

[eddy]

New Build:
CPU - i7-8700
MB - MSI Z390-A Pro

Click to shrink...

It all comes down to price. Are you buying used parts? Because that's the only way that build makes sense to me. There's a reason everyone are building Ryzen 3600 builds right now.

 

[AnimeJesus]

It all comes down to price. Are you buying used parts? Because that's the only way that build makes sense to me. There's a reason everyone are building Ryzen 3600 builds right now.

Click to shrink...

Yes...I'm getting a good deal on the cpu. I've been out of the loop for a while, mind explaining the deal with AMD cpus being popular now. Is it because that's what consoles are using?

 

[eddy]

Yes...I'm getting a good deal on the cpu. I've been out of the loop for a while, mind explaining the deal with AMD cpus being popular now. Is it because that's what consoles are using?

Click to shrink...

AMD delivers the best price/performance in almost every price category at the moment, and the platform isn't a complete dead-end.

The 8700K is a fine CPU and will be a huge upgrade.

 

[AnimeJesus]

AMD delivers the best price/performance in almost every price category at the moment, and the platform isn't a complete dead-end.

The 8700K is a fine CPU and will be a huge upgrade.

Click to shrink...

Cool...thanks for your input.

 

[catpurrcat]

Since we're sharing - build just completed.

 

[パンダさん]

Cant you just buy another Y-cable for more fans?

Click to shrink...

Yes, I've connected 3 fans with 2 y-splitters.

 

[maped]

Question: Can you RMA memory if it's unable to run stable at XMP speeds? I keep getting BSODs at random times exactly like my old computer. Bought a USB stick and tried a memtest and I got 2 errors on test 6. Nothing else is overclocked just the ram at it's xmp profile.

DDR4 just sucks. :/

Click to shrink...

It would help to know your setup, but if it's the NZXT build 9900k and 4000Mhz ram -one, there are no guarantees the XMP-profile just works without a great deal of tuning or at all. Intel rates their k-processor memory controllers at 2666MHz, anything above that is overclocking and not guaranteed, to work or otherwise. Same with motherboards and RAM, QVL-lists and memory manufacturer's specs guarantee that their product is capable of achieving the rated frequencies and timings at optimal circumstances, not that every setup is capable of it.

Was the PC sold with XMP activated, with a promise the 4000MHz RAM speed works? If so, RMA the whole thing. Does memtest error with ram at stock settings? If so, RMA it. If not, raise the clocks a bit at a time to see that they are stable, there's pretty much no point in going beyond 3000Mhz on Intel-platform anyway. Your earlier posts mentioned bluescreening often when overclocking, it's possible that corrupted your install and is also causing troubles so clean re-install might be in order. That said, the ram manufacturer might just RMA it anyways if you ask, but it's not going to change the situation or make it just work with the new RAM.

 

[Allietraa]

Ok Era I've got a very odd problem with my PC. Bought a Ryzen 5 1600 AF and an Asus Prime X470 Pro in early Jan. Had a bunch of problems with the mobo(the bad fan/temp monitoring chip being the key issue but lackluster RAM support too) so I swapped it for a MSI B450 Tomahawk Max. Everything was fine for a while, albeit my CPU temps were a little high(on both mobos) but I was doing light overclocking and messing with fan speeds etc so I didnt think much of it, especially since they were within what I assumed would be an acceptable range(max load temps were about 80-90C under p95 small fft, 75-80C in cinebench r20).
Fast forward a couple weeks and I start getting crazy high temp spikes(100C+, sometimes causing my PC to shutdown) so I assume its a problem with my cooler(which was a custom loop but the pump had been acting up), so I swapped it for a Noctua NH-D15. Temps were a bit higher across the board but I didnt see any 100C+ spikes so I thought it was fixed. After a couple days it goes back to those crazy high spikes and now my idle temps are extremely high too(60-70C on stock voltage with 100% fan speed) and running a stress test gets me right up to 95-100C, during which time i STILL get temp spikes over 100C regardless of what cooler I'm using(tried my loop with an ek mx supremacy evo, nh-d15 and the stock AMD cooler).

I've checked every single thing I can think of. I have taken everything apart and double/triple checked all the cables. Torn apart my loop probably a dozen times. I've remounted the CPU in the socket and checked the pins and the socket itself. I've tried 4 different tubes of thermal paste(two different tubes of nt-h1, nt-h2 and ek ectotherm). I've tried every possible thermal paste application method imaginable. I've tried locking voltages and clock speeds in the BIOS. I've reset the BIOS, updated the BIOS, cleared CMOS, etc and nothing makes any damn difference. I've checked the mounting hardware and it all checks out. I've checked the spread of thermal paste on the CPU/cooler when I remove it and it looks like it should be fine. All the software sensors are reading normal in HWinfo and Ryzen Master. The ambient temp is 22C-ish and my case is an open air Thermaltake Core P5.

I'm at a loss for why this is happening beside some kind of defect in the CPU or the mobo(although why it waited 2-3 weeks to start up I dont understand), and I dont have any other AM4 parts laying around to swap around and check. I've already contacted AMD support but I figured I'd throw this out and see if anyone else has any clues or suggestions.

 

[TaySan]

It would help to know your setup, but if it's the NZXT build 9900k and 4000Mhz ram -one, there are no guarantees the XMP-profile just works without a great deal of tuning or at all. Intel rates their k-processor memory controllers at 2666MHz, anything above that is overclocking and not guaranteed, to work or otherwise. Same with motherboards and RAM, QVL-lists and memory manufacturer's specs guarantee that their product is capable of achieving the rated frequencies and timings at optimal circumstances, not that every setup is capable of it.

Was the PC sold with XMP activated, with a promise the 4000MHz RAM speed works? If so, RMA the whole thing. Does memtest error with ram at stock settings? If so, RMA it. If not, raise the clocks a bit at a time to see that they are stable, there's pretty much no point in going beyond 3000Mhz on Intel-platform anyway. Your earlier posts mentioned bluescreening often when overclocking, it's possible that corrupted your install and is also causing troubles so clean re-install might be in order. That said, the ram manufacturer might just RMA it anyways if you ask, but it's not going to change the situation or make it just work with the new RAM.

Click to shrink...

I'm going test at stock and go from there, but I just hate feeling like I'm not getting the product I paid for even if it is minimal in performance. It advertises itself as 4000mhz.
Here's my setup.

CASE - NZXT H710i (White)

CPU COOLING - NZXT Kraken X72

FAN - AER F 140 FAN (Grey) (3)

FAN - NZXT Aer RGB 2 120mm Fan (RGB)

MOTHERBOARD - ASUS ROG Maximus XI Hero (Wi-Fi)

GPU(GRAPHICS) - NVIDIA GeForce RTX 2080 SUPER - MSI VENTUS XS OC

CPU - Intel Core i9-9900K 8-Core 3.6 GHz

POWER SUPPLY - NZXT E850 Gold Digital PSU

RAM - Team T-FORCE XCALIBUR RGB 16GB (2 x 8GB) 4000MHz (2)

SSD - Samsung 970 EVO Plus (1.0 TB)

EXTRA - NZXT USB Expansion

SERVICES - Standard

SOFTWARE - Microsoft Windows 10 Home

 

[low-G]

Question: Can you RMA memory if it's unable to run stable at XMP speeds? I keep getting BSODs at random times exactly like my old computer. Bought a USB stick and tried a memtest and I got 2 errors on test 6. Nothing else is overclocked just the ram at it's xmp profile.

DDR4 just sucks. :/

Click to shrink...

The only time I personally have had this problem was with crappy BIOSes. I had an Asus board where certain BIOS revisions were stable with my RAM and some BIOS versions weren't. First BIOS didn't properly support, 2nd did, 3rd didn't, 4th finally did again. Now I don't update my BIOS at all lol.

 

[TaySan]

The only time I personally have had this problem was with crappy BIOSes. I had an Asus board where certain BIOS revisions were stable with my RAM and some BIOS versions weren't. First BIOS didn't properly support, 2nd did, 3rd didn't, 4th finally did again. Now I don't update my BIOS at all lol.

Click to shrink...

My last build and this one are both Asus boards :( I think after this I'm done buying anything Asus anymore I don't think I'm even pushing the CPU or ram at all and it seems I can never get anything stable on them. This feels like total bullshit where everyone else like have no problems at all.

 

[ObbyDent]

Is there an all encompassing guide on all the terms and what to look for when theorycrafting? Just reading this thread makes me feel like I'm back in college trying to learn calculus.

 

[GameAddict411]

Ok Era I've got a very odd problem with my PC. Bought a Ryzen 5 1600 AF and an Asus Prime X470 Pro in early Jan. Had a bunch of problems with the mobo(the bad fan/temp monitoring chip being the key issue but lackluster RAM support too) so I swapped it for a MSI B450 Tomahawk Max. Everything was fine for a while, albeit my CPU temps were a little high(on both mobos) but I was doing light overclocking and messing with fan speeds etc so I didnt think much of it, especially since they were within what I assumed would be an acceptable range(max load temps were about 80-90C under p95 small fft, 75-80C in cinebench r20).
Fast forward a couple weeks and I start getting crazy high temp spikes(100C+, sometimes causing my PC to shutdown) so I assume its a problem with my cooler(which was a custom loop but the pump had been acting up), so I swapped it for a Noctua NH-D15. Temps were a bit higher across the board but I didnt see any 100C+ spikes so I thought it was fixed. After a couple days it goes back to those crazy high spikes and now my idle temps are extremely high too(60-70C on stock voltage with 100% fan speed) and running a stress test gets me right up to 95-100C, during which time i STILL get temp spikes over 100C regardless of what cooler I'm using(tried my loop with an ek mx supremacy evo, nh-d15 and the stock AMD cooler).

I've checked every single thing I can think of. I have taken everything apart and double/triple checked all the cables. Torn apart my loop probably a dozen times. I've remounted the CPU in the socket and checked the pins and the socket itself. I've tried 4 different tubes of thermal paste(two different tubes of nt-h1, nt-h2 and ek ectotherm). I've tried every possible thermal paste application method imaginable. I've tried locking voltages and clock speeds in the BIOS. I've reset the BIOS, updated the BIOS, cleared CMOS, etc and nothing makes any damn difference. I've checked the mounting hardware and it all checks out. I've checked the spread of thermal paste on the CPU/cooler when I remove it and it looks like it should be fine. All the software sensors are reading normal in HWinfo and Ryzen Master. The ambient temp is 22C-ish and my case is an open air Thermaltake Core P5.

I'm at a loss for why this is happening beside some kind of defect in the CPU or the mobo(although why it waited 2-3 weeks to start up I dont understand), and I dont have any other AM4 parts laying around to swap around and check. I've already contacted AMD support but I figured I'd throw this out and see if anyone else has any clues or suggestions.

Click to shrink...

Did you run any monitoring software to check the voltage and frequency on windows? My only theory is that something is preventing a proper seal with the cooler even when it appears to be fine. Another thing is that the motherboard could be pushing too much voltage thought CPU because the motherboard is faulty. Or the sensor on the cpu is faulty. The latter two are less likely. In other words many things could be the cause unfortunately.

 

[Dave.]

My last build and this one are both Asus boards :( I think after this I'm done buying anything Asus anymore I don't think I'm even pushing the CPU or ram at all and it seems I can never get anything stable on them. This feels like total bullshit where everyone else like have no problems at all.

Click to shrink...

It's just bad RAM, happens all the time. There is no need to blame ASUS, bad BIOSes, or the whole of DDR4 (lol).

This is why everyone who builds a PC should run memtest86 immediately, preferably before it goes in a case. If the RAM passes at stock but gives errors at it's XMP settings, it is faulty and should be RMA'd. If it gives errors at stock it's obviously faulty by anyone's standards and should be RMA'd. It may be different in USA I guess, but here in UK "Fails memtest86 at XMP settings in an otherwise known good system" is a no-questions-asked insta-RMA. If you are going to try and troubleshoot, first thing to do is get a full run of passes at stock speeds.

Now, there is a tiny chance your CPU's memory controllers can't manage that speed - and I suppose that is perhaps quite a bit higher at 4000MHz. You can test this by taking out one stick and hopefully getting a pass, then using that same stick in the other channel and again getting a pass. But having bought a prebuilt that advertises DDR4000 this is not really your concern. They should have memtest86'd it before it left the factory, obviously didn't. They should now be making good on their promises.

Don't mention to them your overclocking of CPU and GPU, lol 🤐

 

[SmartWaffles]

Ok Era I've got a very odd problem with my PC. Bought a Ryzen 5 1600 AF and an Asus Prime X470 Pro in early Jan. Had a bunch of problems with the mobo(the bad fan/temp monitoring chip being the key issue but lackluster RAM support too) so I swapped it for a MSI B450 Tomahawk Max. Everything was fine for a while, albeit my CPU temps were a little high(on both mobos) but I was doing light overclocking and messing with fan speeds etc so I didnt think much of it, especially since they were within what I assumed would be an acceptable range(max load temps were about 80-90C under p95 small fft, 75-80C in cinebench r20).
Fast forward a couple weeks and I start getting crazy high temp spikes(100C+, sometimes causing my PC to shutdown) so I assume its a problem with my cooler(which was a custom loop but the pump had been acting up), so I swapped it for a Noctua NH-D15. Temps were a bit higher across the board but I didnt see any 100C+ spikes so I thought it was fixed. After a couple days it goes back to those crazy high spikes and now my idle temps are extremely high too(60-70C on stock voltage with 100% fan speed) and running a stress test gets me right up to 95-100C, during which time i STILL get temp spikes over 100C regardless of what cooler I'm using(tried my loop with an ek mx supremacy evo, nh-d15 and the stock AMD cooler).

I've checked every single thing I can think of. I have taken everything apart and double/triple checked all the cables. Torn apart my loop probably a dozen times. I've remounted the CPU in the socket and checked the pins and the socket itself. I've tried 4 different tubes of thermal paste(two different tubes of nt-h1, nt-h2 and ek ectotherm). I've tried every possible thermal paste application method imaginable. I've tried locking voltages and clock speeds in the BIOS. I've reset the BIOS, updated the BIOS, cleared CMOS, etc and nothing makes any damn difference. I've checked the mounting hardware and it all checks out. I've checked the spread of thermal paste on the CPU/cooler when I remove it and it looks like it should be fine. All the software sensors are reading normal in HWinfo and Ryzen Master. The ambient temp is 22C-ish and my case is an open air Thermaltake Core P5.

I'm at a loss for why this is happening beside some kind of defect in the CPU or the mobo(although why it waited 2-3 weeks to start up I dont understand), and I dont have any other AM4 parts laying around to swap around and check. I've already contacted AMD support but I figured I'd throw this out and see if anyone else has any clues or suggestions.

Click to shrink...

Yeah that sounds like a busted CPU, or a busted motherboard sensor. There is no way a normal 1600 goes that high under whatever cooling you have. Also why did you buy a 1600 of all things this year?

 

[Dave.]

Yeah that sounds like a busted CPU, or a busted motherboard sensor. There is no way a normal 1600 goes that high under whatever cooling you have. Also why did you buy a 1600 of all things this year?

Click to shrink...

The 1600 AF is new, it's really a rebadged 2600 or something like that, and it's absurdly cheap.

 

[maped]

It's just bad RAM, happens all the time. There is no need to blame ASUS, bad BIOSes, or the whole of DDR4 (lol).

This is why everyone who builds a PC should run memtest86 immediately, preferably before it goes in a case. If the RAM passes at stock but gives errors at it's XMP settings, it is faulty and should be RMA'd. If it gives errors at stock it's obviously faulty by anyone's standards and should be RMA'd. It may be different in USA I guess, but here in UK "Fails memtest86 at XMP settings in an otherwise known good system" is a no-questions-asked insta-RMA. If you are going to try and troubleshoot, first thing to do is get a full run of passes at stock speeds.

Now, there is a tiny chance your CPU's memory controllers can't manage that speed - and I suppose that is perhaps quite a bit higher at 4000MHz. You can test this by taking out one stick and hopefully getting a pass, then using that same stick in the other channel and again getting a pass. But having bought a prebuilt that advertises DDR4000 this is not really your concern. They should have memtest86'd it before it left the factory, obviously didn't. They should now be making good on their promises.

Don't mention to them your overclocking of CPU and GPU, lol 🤐

Click to shrink...

Again, I'm not saying the manufacturer won't replace the memory out of courtesy, just that it's not guaranteed to fix anything. If the rest of the parts aren't capable, separately or as a combination, it's just not gonna happen. 4000Mhz RAM on 9900k is enough that not all boards and every cpu are going to be able to do it without extensive tweaking or at all. A couple of reviews and user threads I checked on the motherboard had people having to turn down their memory from 4000MHz to be stable.

I'm going test at stock and go from there, but I just hate feeling like I'm not getting the product I paid for even if it is minimal in performance. It advertises itself as 4000mhz.

Click to shrink...

Since it is a ready built machine, the manufacturer is the first place to contact. It's hard to quickly decipher from NZXT's site what their actual stances toward overclocking are and if they overclock the memory as part of the build and guarantee it works, but I'd hazard a guess they don't. Won't hurt asking. And again, if turning down the frequency a bit makes it stable, the loss in performance really is negligible, at worst it's going to cost a frame or two at 1080p. HWUnboxed actually tested this.

 

[SpaceBandito]

Need help! I installed cable extensions this weekend. In the process I unplugged my western digital SDD from one of the SATA memory slots on my MOBO. I accidentally plugged it back into the wrong one it was initially plugged into. I noticed this when I booted up windows after installing the cable extensions. I turned my computer off and plugged my SSD back into its original port on the MOBO.

Please note that Windows is installed on my M.2.

Now my Western Digital takes an extremely long time to load and some applications won't start for an awfully long time. Some applications won't even start at all.

Does anyone have any suggestions?

 

[ploonkey]

Ok Era I've got a very odd problem with my PC. Bought a Ryzen 5 1600 AF and an Asus Prime X470 Pro in early Jan. Had a bunch of problems with the mobo(the bad fan/temp monitoring chip being the key issue but lackluster RAM support too) so I swapped it for a MSI B450 Tomahawk Max. Everything was fine for a while, albeit my CPU temps were a little high(on both mobos) but I was doing light overclocking and messing with fan speeds etc so I didnt think much of it, especially since they were within what I assumed would be an acceptable range(max load temps were about 80-90C under p95 small fft, 75-80C in cinebench r20).
Fast forward a couple weeks and I start getting crazy high temp spikes(100C+, sometimes causing my PC to shutdown) so I assume its a problem with my cooler(which was a custom loop but the pump had been acting up), so I swapped it for a Noctua NH-D15. Temps were a bit higher across the board but I didnt see any 100C+ spikes so I thought it was fixed. After a couple days it goes back to those crazy high spikes and now my idle temps are extremely high too(60-70C on stock voltage with 100% fan speed) and running a stress test gets me right up to 95-100C, during which time i STILL get temp spikes over 100C regardless of what cooler I'm using(tried my loop with an ek mx supremacy evo, nh-d15 and the stock AMD cooler).

I've checked every single thing I can think of. I have taken everything apart and double/triple checked all the cables. Torn apart my loop probably a dozen times. I've remounted the CPU in the socket and checked the pins and the socket itself. I've tried 4 different tubes of thermal paste(two different tubes of nt-h1, nt-h2 and ek ectotherm). I've tried every possible thermal paste application method imaginable. I've tried locking voltages and clock speeds in the BIOS. I've reset the BIOS, updated the BIOS, cleared CMOS, etc and nothing makes any damn difference. I've checked the mounting hardware and it all checks out. I've checked the spread of thermal paste on the CPU/cooler when I remove it and it looks like it should be fine. All the software sensors are reading normal in HWinfo and Ryzen Master. The ambient temp is 22C-ish and my case is an open air Thermaltake Core P5.

I'm at a loss for why this is happening beside some kind of defect in the CPU or the mobo(although why it waited 2-3 weeks to start up I dont understand), and I dont have any other AM4 parts laying around to swap around and check. I've already contacted AMD support but I figured I'd throw this out and see if anyone else has any clues or suggestions.

Click to shrink...

Either there's something wrong with the CPU, or something is trapping heat. Shot in the dark but---did you inspect the CPU pins and socket for any rogue thermal paste?

 

[TC McQueen]

Fast forward a couple weeks and I start getting crazy high temp spikes(100C+, sometimes causing my PC to shutdown) so I assume its a problem with my cooler(which was a custom loop but the pump had been acting up), so I swapped it for a Noctua NH-D15. Temps were a bit higher across the board but I didnt see any 100C+ spikes so I thought it was fixed. After a couple days it goes back to those crazy high spikes and now my idle temps are extremely high too(60-70C on stock voltage with 100% fan speed) and running a stress test gets me right up to 95-100C, during which time i STILL get temp spikes over 100C regardless of what cooler I'm using(tried my loop with an ek mx supremacy evo, nh-d15 and the stock AMD cooler).

I've checked every single thing I can think of. I have taken everything apart and double/triple checked all the cables. Torn apart my loop probably a dozen times. I've remounted the CPU in the socket and checked the pins and the socket itself. I've tried 4 different tubes of thermal paste(two different tubes of nt-h1, nt-h2 and ek ectotherm). I've tried every possible thermal paste application method imaginable. I've tried locking voltages and clock speeds in the BIOS. I've reset the BIOS, updated the BIOS, cleared CMOS, etc and nothing makes any damn difference. I've checked the mounting hardware and it all checks out. I've checked the spread of thermal paste on the CPU/cooler when I remove it and it looks like it should be fine. All the software sensors are reading normal in HWinfo and Ryzen Master. The ambient temp is 22C-ish and my case is an open air Thermaltake Core P5.

I'm at a loss for why this is happening beside some kind of defect in the CPU or the mobo(although why it waited 2-3 weeks to start up I dont understand), and I dont have any other AM4 parts laying around to swap around and check. I've already contacted AMD support but I figured I'd throw this out and see if anyone else has any clues or suggestions.

Click to shrink...

Have you dialed down Load Line Calibration to level 3 or lower? The mobo could be pumping too much voltage through it too quickly to maintain frequency.

 

[sleepnaught]

So, HL: Alyx is releasing in a couple weeks and all I have is this puny 1070 to power my Valve Index. Obviously, this isn't going to cut it. I was banking on the GTX 3000 series card to be out by now when I purchased my Index a couple months back. They're still MIA. It looks like I can get $175-$200 for my 1070 still. I'm thinking about selling it and then just shelling out $300ish for a 2070 Super, this way I can at least have a pretty good experience with HL: Alyx, while not investing too much on an old card.

 

[TaySan]

I tested each stick individually with the xmp enabled and each of them did a full 4 passes with no errors. Now I"m testing them all in at stock settings and no errors so far. Really weird I got 2 errors within the first ten minutes when I did the first test so why everything is fine now?

Going to leave it on all night.

 

[Crazymoogle]

Need help! I installed cable extensions this weekend. In the process I unplugged my western digital SDD from one of the SATA memory slots on my MOBO. I accidentally plugged it back into the wrong one it was initially plugged into. I noticed this when I booted up windows after installing the cable extensions. I turned my computer off and plugged my SSD back into its original port on the MOBO.

Please note that Windows is installed on my M.2.

Now my Western Digital takes an extremely long time to load and some applications won't start for an awfully long time. Some applications won't even start at all.

Does anyone have any suggestions?

Click to shrink...

Sorry, your writing suggests a few different scenarios (since "SATA Memory Slot" could mean SATA Port or M.2 SATA Port):
s1: SSD is M2 SATA, put it in the wrong M2 slot (as in: you moved your boot drive)
s2: SSD is 2.5" SATA, attached it to the wrong SATA port (as in: you moved your secondary drive)

If Scenario 2:
Step 1: Unplug the WD drive and see if performance goes back to normal
Step 2: Verify that M2 is first on your boot order in BIOS/UEFI
Step 3: Try connecting the drive to a different SATA port (and re-seat the cable on both ends)
Step 4: Try a different SATA cable

Beyond that...what are the cable extensions for? Is the drive letter the same? (It's possible some windows registry messup happened)

 

[maped]

I tested each stick individually with the xmp enabled and each of them did a full 4 passes with no errors. Now I"m testing them all in at stock settings and no errors so far. Really weird I got 2 errors within the first ten minutes when I did the first test so why everything is fine now?

Going to leave it on all night.

Click to shrink...

Two sticks strain the memory controller more than one, might be that that's just the limit. If the sticks individually pass with the XMP enabled, it would seem it's either the processor or the board that's holding you back. It might be just a case a bumbing the appropriate voltages up a bit to get it to work but my Intel setups are from DDR3-era so can't advise more on that.

 

[Kiyoshi]

Having a right battle with my PC at the moment... GPU driver crashes or hard locks after any short period of gaming time :( It's not a new system, and was running fine for months, so this has just started happening.

I have a Ryzen 2700X, GTX 1070, 32GB RAM

Temperatures across the whole system are fine. I've tried the following:-

• Disabled all non-essential processes
• Clean driver re-install (safe mode & DDU method)
• Regressed to an older driver version (two iterations)
• Re-seated GPU
• Registry hack to add a TdrLevel key
• Disabled CPU overclocks
• Fiddled in MSI afterburner
  • Increased power limited
  • Decreased power limit
  • Lowered core clock to lowest amount
  • Lowered memory clock to lowest amount

Anything else I could try, short of reinstalling Windows?

 

[GamingRobioto]

I'm on the edge of buying my components for my next build and I'm still on the fence about whether to get a i7 9700k or pay the extra £125 for the i9 9900k.

Is the i9 9900k overkill for gaming? Bear in mind that this system will be built to last me 5-6 years with only a GPU upgrade if it's necessary. (I've had my current i7 4790k build since 2014). Or am I perfectly fine with the i7 9700k?

Also if I go for the the i7 9700k I will go with this AIO cooler (Corsair H100i): https://www.scan.co.uk/products/cor...e-hydro-cpu-cooler-2x-120mm-ml-pro-rgb-fans-r

If I do decide to go with the i9 9900k will this dual radiator be sufficient? Or would it be better to get a triple for overclocking (H150i): https://www.scan.co.uk/products/cor...ro-cpu-cooler-3x-120mm-ml-series-fans-rgb-pum

If I do need a triple radiator I'll probably have to rethink the airflow in my case, but that's fine. Also kind of what RGB on the fans which isn't on the H150i, but isn't the end of he world, just trying to make my build look amazing as that is part of the fun. :)

Thanks for any help.

 

[macindc]

Anything else I could try, short of reinstalling Windows?

Click to shrink...

I hate to say it, but considering you're saying this just started happening, and if there's no clear change (OS update, driver update, hardware change) that instigated it, your GPU could actually be just starting to die.

I'd check to see if it's still under warranty (a 1070 isn't THAT old and many manufacturers go 3-5 years) and reach out about a possible RMA.

I'm on the edge of buying my components for my next build and I'm still on the fence about whether to get a i7 9700k or pay the extra £125 for the i9 9900k.

Is the i9 9900k overkill for gaming? Bear in mind that this system will be built to last me 5-6 years with only a GPU upgrade if it's necessary. (I've had my current i7 4790k build since 2014). Or am I perfectly fine with the i7 9700k?

Click to shrink...

YES an i9 is overkill for gaming. If you're also doing intense CPU work for other things like video editing or computational work go for it but for a gaming rig pretty much anything with 4+ cores on the market right now is adequate. It's the GPU that's more important.

 

[GamingRobioto]

I hate to say it, but considering you're saying this just started happening, and if there's no clear change (OS update, driver update, hardware change) that instigated it, your GPU could actually be just starting to die.

I'd check to see if it's still under warranty (a 1070 isn't THAT old and many manufacturers go 3-5 years) and reach out about a possible RMA.




YES an i9 is overkill for gaming. If you're also doing intense CPU work for other things like video editing or computational work go for it but for a gaming rig pretty much anything with 4+ cores on the market right now is adequate. It's the GPU that's more important.

Click to shrink...

Thanks, I will not be upgrading my GPU until Nvidia release the next generation of GPUs, which will hopefully be before Cyberpunk 2077 in September. So for now I am going to keep going with my 980ti while building everything else this month. So I'm not budgeting at all for a GPU yet. I have around £1,200 and effectively need a case, CPU, Cooler, Motherboard, Memory and a SSD. I will be reusing my PSU and 3 SSDs from my current build.

I think I'm just going to go with the i7 and save myself a couple of hundred pounds.

 

[ShaggsMagoo]

What are we thinking will be the odds of the next AMD generation of CPUs and the new Nvidia GPUs dropping before Cyberpunk? My 980ti and 6700k are holding up pretty well for 1080p, but start showing their age when I try 1440p on newer games.

 

[Crazymoogle]

Thanks, I will not be upgrading my GPU until Nvidia release the next generation of GPUs, which will hopefully be before Cyberpunk 2077 in September. So for now I am going to keep going with my 980ti while building everything else this month. So I'm not budgeting at all for a GPU yet. I have around £1,200 and effectively need a case, CPU, Cooler, Motherboard, Memory and a SSD. I will be reusing my PSU and 3 SSDs from my current build.

I think I'm just going to go with the i7 and save myself a couple of hundred pounds.

Click to shrink...

I would agree that you'd be better off saving budget for (a) the GPU, (b) the GPU...etc. It's really the most important part of your build. You're way better off overspending on GPU and sticking with the 9700K CPU, which is already a bit of a beast in general. If nothing else this gives you the breathing room to pick out exactly the motherboard you want.

The H100i is kind of the tired champion of AIO but it's also had so many revs that it's the most polished. The Platinum loads up on RGB and has a zero RPM mode in Corsair's software, which I have to admit would have made my - 6 year old - H100i immensely more palatable. Personally I prefer 2x140mm radiators now, but whether you go that or 3x120 you have to better consider the case you want. Meshify C and P400A are possibly the most flexible, but lots of people make NZXT builds work, so...

What are we thinking will be the odds of the next AMD generation of CPUs and the new Nvidia GPUs dropping before Cyberpunk? My 980ti and 6700k are holding up pretty well for 1080p, but start showing their age when I try 1440p on newer games.

Click to shrink...

AMD 4000: 33%? It really smells of being a Q4 product right now. It's possible though, sure.
RTX 3000: 99% will be out before September 17...I'd be more surprised if it wasn't out by May. Of course, that makes the 2080ti contest a bit awkward, but that hasn't stopped companies before...

 

[GamingRobioto]

I would agree that you'd be better off saving budget for (a) the GPU, (b) the GPU...etc. It's really the most important part of your build. You're way better off overspending on GPU and sticking with the 9700K CPU, which is already a bit of a beast in general. If nothing else this gives you the breathing room to pick out exactly the motherboard you want.

The H100i is kind of the tired champion of AIO but it's also had so many revs that it's the most polished. The Platinum loads up on RGB and has a zero RPM mode in Corsair's software, which I have to admit would have made my - 6 year old - H100i immensely more palatable. Personally I prefer 2x140mm radiators now, but whether you go that or 3x120 you have to better consider the case you want. Meshify C and P400A are possibly the most flexible, but lots of people make NZXT builds work, so...



AMD 4000: 33%? It really smells of being a Q4 product right now. It's possible though, sure.
RTX 3000: 99% will be out before September 17...I'd be more surprised if it wasn't out by May. Of course, that makes the 2080ti contest a bit awkward, but that hasn't stopped companies before...

Click to shrink...

Thank you, I assume you aren't a fan of NZXT? How come? I really like their aesthetically cases so I'm going for a H700, my current build has a H440. I'm avoiding the i versions as the CAM smart control think sounds horrendous.

What alternative would you recommend to the H100i? Or do you think that's a good choice.

 

[Crazymoogle]

Thank you, I assume you aren't a fan of NZXT? How come? I really like their aesthetically cases so I'm going for a H700, my current build has a H440. I'm avoiding the i versions as the CAM smart control think sounds horrendous.

What alternative would you recommend to the H100i? Or do you think that's a good choice.

Click to shrink...

CAM is terrible but their cases are slick, no doubt. Arguably the best looking cases out there? But I'm not a fan of how they block off the front. It's kind of a waste for RGB and a nightmare for airflow. So for now, the P400A and Meshify C are way better options to me. Or even the Cooler Master Case with the giantic 2x200m front fans (H500M?) The P400A even comes with front RGB fans in the box (not great ones by any means, but functional).

I ended up going with the Meshify C on my recent build, mainly because the front mesh holds 3x120 or 2x140, and then 2x120 on the top. The P400A is the king of airflow but doesn't use a dust filter, so...depends on your climate. The spider-panel on the front is not ideal for RGB, but it still shines through if your fans are good.

I think the H100i Platinum is great, like I said, it's a well rehearsed design. I think the NZXT Kraken models with the front LCD are more popular these days? I gave up on pump noise and went with an air cooler though...

 

[Deleted member 2172]

So I'm not budgeting at all for a GPU yet. I have around £1,200 and effectively need a case, CPU, Cooler, Motherboard, Memory and a SSD. I will be reusing my PSU and 3 SSDs from my current build.

Click to shrink...

Thats a great budget for those parts. You'll likely be able to get everything you need there for less than £1k.

 

[パンダさん]

I'm going to sell my 9700k mITX build to a close friend, because it's exactly what he wants.
I'll be doing a new one from scratch with an ATX MB and 3900x, because I need more threads and less noise.
I'll keep my 2 SATA SSDs as well as my RTX 2070 Super.

The only thing I'm not sure about is the PSU.
The board has 2 CPU Power connectors, but as I could find out, you can use it with only one connected.
The be Quiet! Straight Power 11 650W does only have one CPU PWR.
The Straight Power 11 750W has two of them.
But imo 650W is more than enough.
The 650W Platinum costs the same as the 750W Gold.

I'm also thinking about throwing in a 3700x, upgrade to a 4900x later, when it's out, and replace the 2600x of my GF with the 3700x.

What dou you think?

https://de.pcpartpicker.com/user/febLey/saved/zMx9NG

 

[GamingRobioto]

CAM is terrible but their cases are slick, no doubt. Arguably the best looking cases out there? But I'm not a fan of how they block off the front. It's kind of a waste for RGB and a nightmare for airflow. So for now, the P400A and Meshify C are way better options to me. Or even the Cooler Master Case with the giantic 2x200m front fans (H500M?) The P400A even comes with front RGB fans in the box (not great ones by any means, but functional).

I ended up going with the Meshify C on my recent build, mainly because the front mesh holds 3x120 or 2x140, and then 2x120 on the top. The P400A is the king of airflow but doesn't use a dust filter, so...depends on your climate. The spider-panel on the front is not ideal for RGB, but it still shines through if your fans are good.

I think the H100i Platinum is great, like I said, it's a well rehearsed design. I think the NZXT Kraken models with the front LCD are more popular these days? I gave up on pump noise and went with an air cooler though...

Click to shrink...

Thanks, I'll take a look at those cases, I'm going to order on Friday so I can build over the weekend so I have time to consider what I'm going to do.

I'm doing a blue themed build this time around, the blue/black H700 first this nicely.

 

[Kiyoshi]

Having a right battle with my PC at the moment... GPU driver crashes or hard locks after any short period of gaming time :( It's not a new system, and was running fine for months, so this has just started happening.

I have a Ryzen 2700X, GTX 1070, 32GB RAM

Temperatures across the whole system are fine. I've tried the following:-

• Disabled all non-essential processes
• Clean driver re-install (safe mode & DDU method)
• Regressed to an older driver version (two iterations)
• Re-seated GPU
• Registry hack to add a TdrLevel key
• Disabled CPU overclocks
• Fiddled in MSI afterburner
  • Increased power limited
  • Decreased power limit
  • Lowered core clock to lowest amount
  • Lowered memory clock to lowest amount

Anything else I could try, short of reinstalling Windows?

Click to shrink...

I hate to say it, but considering you're saying this just started happening, and if there's no clear change (OS update, driver update, hardware change) that instigated it, your GPU could actually be just starting to die.

I'd check to see if it's still under warranty (a 1070 isn't THAT old and many manufacturers go 3-5 years) and reach out about a possible RMA.

Click to shrink...

I think I found the answer! I'm using an NVMe drive and I read that my BIOS should have Windows WHQL mode enabled, but it was disabled. I remember having a play around with the BIOS settings a while back (around the time I started experiencing the issues), so I've turned this on now and I haven't had a single crash since. I had a GPU stress test running (Heaven) for a while, and have tried a few games. I managed to play a couple of hours of KC:D without a crash. So fingers crossed this is resolved now.

 

[Allietraa]

Did you run any monitoring software to check the voltage and frequency on windows? My only theory is that something is preventing a proper seal with the cooler even when it appears to be fine. Another thing is that the motherboard could be pushing too much voltage thought CPU because the motherboard is faulty. Or the sensor on the cpu is faulty. The latter two are less likely. In other words many things could be the cause unfortunately.

Click to shrink...

I checked all the sensors(voltage, current, etc) in HWinfo(on both the CPU itself and the motherboards sensors for those things) and Ryzen Master and they all read what I'd expect(i.e. what i set in the mobo, or within proper ranges for stock/auto).

Either there's something wrong with the CPU, or something is trapping heat. Shot in the dark but---did you inspect the CPU pins and socket for any rogue thermal paste?

Click to shrink...

I've checked it a couple times and I dont see any. Looked over the socket with a light and it looks fine too.

Have you dialed down Load Line Calibration to level 3 or lower? The mobo could be pumping too much voltage through it too quickly to maintain frequency.

Click to shrink...

I didnt try all the steps but IIRC I did go through a few of them and found no diff


I sent in some more info to AMD this morning and added a couple screenshots of temps so hopefully they'll be quick with figuring this out or just let me RMA it lol

 

[Crazymoogle]

The only thing I'm not sure about is the PSU.
The board has 2 CPU Power connectors, but as I could find out, you can use it with only one connected.
The be Quiet! Straight Power 11 650W does only have one CPU PWR.
The Straight Power 11 750W has two of them.
But imo 650W is more than enough.
The 650W Platinum costs the same as the 750W Gold.

I'm also thinking about throwing in a 3700x, upgrade to a 4900x later, when it's out, and replace the 2600x of my GF with the 3700x

Click to shrink...

3700X is a gaming step down from the 9700K in many specs, but you said you needed the threads, which suggests the 3900X is the purpose of your build?

Normally the PSU incompatibility is about an extra 4pin needed for LN2 mode?

The PSU incompatibility is probably to be the usual case of a motherboard having the extra 4pin for LN2 mode, which apparently even 3900x stock won't need. The MEG UNIFY has dual 8pin...but then this thread says you only need one. I'm inclined to believe whatever power recommendation the video link is to Buildzoid who knows wtf he is talking about.

 

[shnurgleton]

Encountering an issue with my louqe ghost s1 build.

Have an aorus x570 mini itx board with a ryzen, looking to use the noctua L12s cooler. I affixed this boy to the board only to learn that it's a few mm too tall for me to slide on the case side plate without significant force. a googling has led me to a number of reddit threads about bending (!) the cooler or using a beltsander (!!!) to shave down the fins on the cooler to get it to fit. I don't have access to a fucking metalworking shop so the beltsander option is not on the table for me. any of yall have experience with this particular setup and have advice?

 

[パンダさん]

3700X is a gaming step down from the 9700K in many specs, but you said you needed the threads, which suggests the 3900X is the purpose of your build?

Click to shrink...

Yes, I'm aware of that, but thanks for pointing it out.
I can definitely live with this.

Normally the PSU incompatibility is about an extra 4pin needed for LN2 mode?

The PSU incompatibility is probably to be the usual case of a motherboard having the extra 4pin for LN2 mode, which apparently even 3900x stock won't need. The MEG UNIFY has dual 8pin...but then this thread says you only need one. I'm inclined to believe whatever power recommendation the video link is to Buildzoid who knows wtf he is talking about.

Click to shrink...

So as I understand, using the 650W PSU shouldn't be a problem.
Thanks you very much for the advice.

 

[Skyfireblaze]

Yes, I'm aware of that, but thanks for pointing it out.
I can definitely live with this.


So as I understand, using the 650W PSU shouldn't be a problem.
Thanks you very much for the advice.

Click to shrink...

As a rule of thumb 600-650w should be able to handle everything consumer-level nowadays bar exotic configurations like SLI/Crossfire, Dual-chip GPUs, multi-socket CPU boards or having something unusual like 10 or more HDDs.