Help me build my PC: 2024 Catch-All

Thanks.

Going to let it run overnight, and tomorrow I will be digging into the bios to see how to turn on the wifi, as it does not appear to exist according to Windows, and I don't know if it is because I need to update the drivers, or if it is just turned off.

*Legion* wrote:

Highly recommend the FakeSpot browser extension, which will help point out bogus sellers on Amazon and eBay. Available for Firefox as well as the popular ad company spyware browsers you shouldn't be using.

Still recommend this extension? Recent reviews have been unkind.

I don't know if it is because I need to update the drivers, or if it is just turned off.

More than likely the former. May need a firmware update too.

Sometimes the network troubleshooter can detect and fix it too.

"Memory training"? Is that a term for formatting? Or something else?

Edit - I see, another word for "initializing", which includes setting up the pool of RAM for the OS to use.

Robear wrote:

"Memory training"? Is that a term for formatting? Or something else?

Memory training is a process where the system tests various timings and parameters on newly installed RAM sticks in order to determine what settings are stable or not. The settings are then saved so that training does not need to be repeated on each subsequent boot (until a change in RAM configuration is detected).

For the user, what memory training looks like is their newly built system taking an exorbitant amount of time to actually POST and show a video signal, letting them spend anywhere from 30 seconds to up to 15 minutes living in abject fear that their expensive new machine is broken.

*Legion* wrote:

HWINFO64 monitors everything. Needs to be combined with RTSS for overlay though, and fiddling with both for overlays is not intuitive. Absolute must-have just for being able to open the app and see ALL the sensors though.

The NVIDIA overlay from GeForce Experience (or, better yet, the new beta NVIDIA app) can monitor GPU temps and some other useful stuff like system latency, but not CPU temps.

I have MSI Afterburner, RTSS and Hwinfo running. I don't do anything with MSI tuner other than I heard IT was necessary for the overlay along with RTSS. The Hotkey to turn on the overlay is set in MSI Tuner. Is there a way to assign that hotkey elsewhere and just stop using MSI Afterburner?

Also @mudbunny, just in case you didn't know, you have to actually turn on a setting in BIOS to get your full DDR6000. On intel boards its called XMP, on AMD boards I think its called EXPO?

Will defer to you when you want to flip that switch, but I would just make sure you have the latest BIOS and Drivers, and then flip it on.. else your DDR6000 is just DDR3000 (or 3400).

Thanks Legion!

Carlbear95 wrote:

Also @mudbunny, just in case you didn't know, you have to actually turn on a setting in BIOS to get your full DDR6000. On intel boards its called XMP, on AMD boards I think its called EXPO?

Will defer to you when you want to flip that switch, but I would just make sure you have the latest BIOS and Drivers, and then flip it on.. else your DDR6000 is just DDR3000 (or 3400).

Thanks!!

I did not know that.

What’s the difference that I would see?

mudbunny wrote:
Carlbear95 wrote:

Also @mudbunny, just in case you didn't know, you have to actually turn on a setting in BIOS to get your full DDR6000. On intel boards its called XMP, on AMD boards I think its called EXPO?

Will defer to you when you want to flip that switch, but I would just make sure you have the latest BIOS and Drivers, and then flip it on.. else your DDR6000 is just DDR3000 (or 3400).

Thanks!!

I did not know that.

What’s the difference that I would see?

Real World? I'm sure someone smarter than me can answer. I think for me its "You paid for DDR-6000, you should get DDR-6000". Understanding that if you're not comfortable with that, and given your recent RAM experiences I wouldn't blame you, that you may not even notice the difference.

Carlbear95 wrote:

I have MSI Afterburner, RTSS and Hwinfo running. I don't do anything with MSI tuner other than I heard IT was necessary for the overlay along with RTSS. The Hotkey to turn on the overlay is set in MSI Tuner. Is there a way to assign that hotkey elsewhere and just stop using MSI Afterburner?

Video on using HWINFO + RTSS without using Afterburner:

Carlbear95 wrote:

Also @mudbunny, just in case you didn't know, you have to actually turn on a setting in BIOS to get your full DDR6000. On intel boards its called XMP, on AMD boards I think its called EXPO?

Will defer to you when you want to flip that switch, but I would just make sure you have the latest BIOS and Drivers, and then flip it on.. else your DDR6000 is just DDR3000 (or 3400).

And some AMD boards will support both. I assume Intel boards might also do the same.

MannishBoy wrote:
Carlbear95 wrote:

Also @mudbunny, just in case you didn't know, you have to actually turn on a setting in BIOS to get your full DDR6000. On intel boards its called XMP, on AMD boards I think its called EXPO?

Will defer to you when you want to flip that switch, but I would just make sure you have the latest BIOS and Drivers, and then flip it on.. else your DDR6000 is just DDR3000 (or 3400).

And some AMD boards will support both. I assume Intel boards might also do the same.

That's interesting. Didn't realize that. I thought XMP and EXPO were trademarked to their respective chip overlords, wouldn't expect them to offer both. I'm assuming there's safeguards so you don't turn both on? My Intel Mobo is barely a year old and only has XMP.

mudbunny wrote:
Carlbear95 wrote:

Also @mudbunny, just in case you didn't know, you have to actually turn on a setting in BIOS to get your full DDR6000. On intel boards its called XMP, on AMD boards I think its called EXPO?

Will defer to you when you want to flip that switch, but I would just make sure you have the latest BIOS and Drivers, and then flip it on.. else your DDR6000 is just DDR3000 (or 3400).

Thanks!!

I did not know that.

What’s the difference that I would see?

Don't forget to check settings every time you update your BIOS. I just updated today and found on my AMD system, XMP/EXPO and PBO were both turned back off.

At least updating BIOS is really easy now. Just go into the BIOS settings at boot and with the usb drive already inserted, selection the option and watch it go.

Carlbear95 wrote:

That's interesting. Didn't realize that. I thought XMP and EXPO were trademarked to their respective chip overlords, wouldn't expect them to offer both. I'm assuming there's safeguards so you don't turn both on? My Intel Mobo is barely a year old and only has XMP.

XMP and EXPO are essentially the exact same thing: protocols for the RAM to communicate predefined configurations that the RAM supports. One's just a proprietary protocol from Intel and the other's an open protocol from AMD that provides almost 100% the exact same data (there are, IIRC, some small timing things in EXPO that aren't available in XMP as XMP is older).

Most RAM sticks provide their profiles over one or the other, so if you try to load up XMP profiles on RAM that only offers its profiles via EXPO, then all you're going to see when you try to select a profile is a blank list.

Some RAM sticks do provide profiles over both XMP and EXPO, and if a board allowed you to "enable both", then I imagine all that would happen is that it would set the RAM's config to the profile selected via one, and then immediately set the RAM's config to the profile selected on the other - which are probably the exact same values (unless you intentionally choose Profile 1 on XMP and Profile 2 on EXPO, for example.) I've never tested this, and I imagine most BIOS UIs probably disable one if you click to enable the other, but even if you were allowed to select both, I expect the RAM would just end up loaded with whichever profile the BIOS loads last.

*Legion* wrote:

Some RAM sticks do provide profiles over both XMP and EXPO, and if a board allowed you to "enable both", then I imagine all that would happen is that it would set the RAM's config to the profile selected via one, and then immediately set the RAM's config to the profile selected on the other - which are probably the exact same values (unless you intentionally choose Profile 1 on XMP and Profile 2 on EXPO, for example.) I've never tested this, and I imagine most BIOS UIs probably disable one if you click to enable the other, but even if you were allowed to select both, I expect the RAM would just end up loaded with whichever profile the BIOS loads last.

My MSI MAG B650 Tomahawk has both profile sets. You pick one or the other and let it configure.

Hey all, having some trouble with games crashing out on me again. Most of the time it happens I'm able to track down some form of workaround, but a few games I've just had to drop outright. Can't be sure, but research indicates it *might* be something in my Nvidia graphics card (RTX 3070), either it running too hot or being faulty. I'm considering upgrading it now, and possibly switching to an AMD card as most of the complaints I'm seeing about these issues in my research are blaming the Nvidia card.

I use the system about equally for office/productivity and gaming. I run three monitors (all via Displayport I believe on my current card), but only at 1920x1080, so no need for 4k or anything. Any suggestions as to what would be a good replacement/upgrade card, and whether switching to AMD is the way to go?

Other possibly relevant stats for the system are below.

Thanks in advance!

Processors: AMD Ryzen 9 5900X 12-Core 3.7GHz (4.8GHz Max Boost)
Motherboard: ASRock X570 Extreme4
Memory: 64GB CORSAIR VENGEANCE RGB RT DDR4 3200MHz (2x32GB)
System Cooling: CORSAIR H150i PRO XT RGB Cooler
Graphics Cards: NVIDIA 8GB GeForce RTX 3070
Power Supply: CORSAIR RM850 SERIES PLUS GOLD PSU

When did the crashes start? Just replacing things, while sometimes effective, is an aggressive approach. If you’re looking to use this as an excuse to upgrade that’s cool, but there’s also no promise that it will fix the problem.

JC wrote:

When did the crashes start? Just replacing things, while sometimes effective, is an aggressive approach. If you’re looking to use this as an excuse to upgrade that’s cool, but there’s also no promise that it will fix the problem.

It's been the better part of a year that I've been having the crashes, over a variety of games. Machine itself is two and a half years old, so I don't need to upgrade at this point, but the crashes have become obnoxious enough that I'm looking at a longer term solution than finding workaround like capping FPS and reducing graphics quality and such.

Have you confirmed via monitoring that it's an actual overheating issue? Is the PSU only 2.5 years old as well?

Have you tried running just one stick of RAM?

There's a ton of troubleshooting that could be done prior to replacing expensive parts.

Teneman wrote:

and possibly switching to an AMD card as most of the complaints I'm seeing about these issues in my research are blaming the Nvidia card.

Unless you can narrow things down to a very specific issue, saying the GPU brand that most gaming PCs use is the problem is probably not very insightful or accurate troubleshooting. I mean, upgrade to an AMD card by all means if that's what you want, but absent any specific verifiable information, this just sounds like people talking out their backsides.

So what exactly are people claiming?

either it running too hot or being faulty

Are you monitoring temperatures?

Most of the time it happens I'm able to track down some form of workaround

What do these workarounds look like?

*Legion* wrote:
either it running too hot or being faulty

Are you monitoring temperatures?

Most of the time it happens I'm able to track down some form of workaround

What do these workarounds look like?

I am monitoring temps, not continuously, but I check them when there's been an issue. They run fairly cool normally (showing GPU at 35 and Hot Spot at 46 degrees right now). When running games the card runs hotter of course, but usually not higher than the low 80s - which supposedly is ok for the RTX 3070.

The workarounds vary by game, but the most frequently successful is capping the FPS of the game at 60, although sometimes I can cap it in the game settings itself, and sometimes I have to cap it with a custom profile for the game within the NVidia Control panel, even if the game has its own settings.

The thing that's become frustrating is this isn't something that only happens on new high end games that you'd expect to be pushing the graphics card. I'm even getting crashes on some older games, games that my wife is playing at the same time with no problems on her older laptop.

I know there's likely more diagnosing and troubleshooting that can be done, I'm just tired of spending half of the little game time I get troubleshooting. I rarely had graphics/gaming issues with my last machine (self built) and seem to have them far too frequently with this one (Origin PC built). I'm just looking for the most likely culprit in the hopes that I can swap that component out and get back to gaming without having to play tech support first.

Are you running any extraneous software? Razer software, Steel Series. Razer software is notoriously bad. You say its a pre-built, have you cleaned out all of the bloatware they may have installed with a tool like Revo?

Also is it just a game crash to windows? Or is there more to it like reboot, blue screen or freeze?

I’d also uninstall all video card drivers, reboot, run driver cleaner, reboot, and then do a clean install of video drivers.

Carlbear is spot on as well. Are these BSODs? If yes, what’s the error message when it crashes.

Origin didn't put too much bloatware on the machine, but I did clean even that off. Logitech G Hub and iCue are the only accessory/gaming software that's in the background, plus Steam of course. There are a few other background apps from the office/productivity side (Veeam, Voice Attack, Dropbox, DisplayFusion), but I've individually disabled those during some of the early crashes to make sure they didn't have any impact.

The crashes are rarely BSOD. It depends on the game, but most often its either a crash to desktop, a crash to desktop with an error message (these are the ones I've researched back and found fingers pointing at NVidia), or just a hard freeze of the game requiring ending the task through task manager. I do keep my drivers up to date, but haven't done a clean install of them for awhile, I'll give that a try.

Teneman wrote:

The workarounds vary by game, but the most frequently successful is capping the FPS of the game at 60, although sometimes I can cap it in the game settings itself, and sometimes I have to cap it with a custom profile for the game within the NVidia Control panel, even if the game has its own settings.

OK, those workarounds sound mostly like what they're doing is load reduction.

I am monitoring temps, not continuously, but I check them when there's been an issue. They run fairly cool normally (showing GPU at 35 and Hot Spot at 46 degrees right now). When running games the card runs hotter of course, but usually not higher than the low 80s - which supposedly is ok for the RTX 3070.

Low 80s is getting on the toasty side. It's below the max operating temperature, which IIRC is 93 degrees for Ampere, but 83 degrees is the default temperature soft limit, which the card actively tries to avoid exceeding. I wonder if your card is really throttling hard in order to try and keep under 83. You also said "usually", which makes me wonder if it's peaking a bit higher than that right before failure.

I wonder, though, if what's happening is less temperature related, and more related to power delivery. Ampere was known to sometimes have some gnarly transient power spikes (although more in the top-end, 3080 and 3090 cards). 850 watts ought to be enough for a system with a 3070 - I ran mine with less - but if the power supply is operating less than optimally, failing to maintain power delivery at max GPU load is a behavior I might expect to see.

One thing I would try is running a stress test like FurMark, and watching a few metrics like a hawk while it's running: GPU temperature, GPU core and memory clocks, GPU power draw, and GPU throttling. Install HWINFO64 if you don't already have it, so that you can easily see values like this:

IMAGE(https://i.imgur.com/qYpbHDI.png)

My suspicion list, from highest to lowest, would be:

* Power delivery issue
* Thermal throttling
* Fault in the card
* Software related issue
* Memory fault
* Other platform fault

Since power delivery is atop on my suspicion list: the 3070 has two PCIe power connectors. Are you connecting those off of the same cable, or two different cables? For higher draw GPUs, I strongly recommend those connections come off of two separate cables.

I don't know what case you're using, but what is the ventilation like? Is it one of those solid front cases that really chokes off intake, or more of a mesh front with good airflow? What does the case fan configuration look like? I see you're running a pretty beefy 360mm AIO, so I'm assuming you're not a total noob in this space.

(Origin PC built)

Oh, it's prebuilt. Origin, though, so not a shoddy slap-it-together operation. Is your support window expired or something? Having someone to solve/fix these issues is the reason to buy from a boutique system integrator like Origin.

The advice of doing a full driver wipe and clean install of GPU drivers is a good one. DDU will help with purging any remnants of old drivers. I also wouldn't rule out doing a full system wipe, I'm pretty sure Origin provides recovery media.

I would expect memory faults to manifest in more ways than just this, but running memtest86+ to try and rule out bad sticks is an easy step to take.

I used to have a Radeon RX5700 XT card that had a similar problem. Random crashes and freezes that sound a lot like the ones you are describing, typically under heavy gaming loads but sometimes even at more normal use loads.

I tried all or most of the stuff Legion mentioned. I had the two PCI-e cables on separate cables, a more than capable PSU, a case with good airflow, normal temperatures. Undervolting the card did nothing. What ended up solving the issue was lowering the maximum power draw of the card by 2-3% in the Radeon software. The drop in performance, if any, was imperceptible and it never crashed again. I'm sure nVidia software allows you to do the same. Maybe that's something you can try.

You could also try looking at the Windows Event Viewer after a freeze or crash to see if you identify any helpful information. If you see an error that pops up every time there is a crash, Google it and you may find other people with the same issue. That's one thing I did to get to the solution to my problem.

The other corollary of my story is that getting rid of your card and getting an AMD card is not guarantee. All hardware has a small chance of these things happening.

Pink Stripes wrote:

What ended up solving the issue was lowering the maximum power draw of the card by 2-3% in the Radeon software. The drop in performance, if any, was imperceptible and it never crashed again. I'm sure nVidia software allows you to do the same. Maybe that's something you can try.

Easiest way to do this is with MSI Afterburner. Power limit and the temperature limit are a couple easy sliders:

IMAGE(https://i.imgur.com/WXxqsUu.png)

My support through Origin is definitely past, one year support and I'm about two and a half years in on the machine.

I downloaded Furmark and HWiNFO64. Ran Furmark and the power usage immediately spiked to 100%, and then bounced between 99% and 101%. On HWiNFO the Power Performance Limit also immediately switched to Yes.

I confess that we're out past the edge of my knowledge at this point, but I assume that these results bear out your suspicion that it's a power delivery issue Legion. I checked the cables, and it appears that there are two separate cables from the card to the PSU. The cables are all nicely routed and tied together, and I suppose it is possible that it is actually one cable disguised as two and I'm just not seeing the junction/split, but I do not think so. So what would the next step be?

Would reducing the max power draw be the next step as per Pink Stripe's suggestion? I didn't see a way to do that in increments in Nvidia Control Panel, but I do have MSI Afterburner so could do it from there if that would be helpful.

Teneman wrote:

Would reducing the max power draw be the next step as per Pink Stripe's suggestion? I didn't see a way to do that in increments in Nvidia Control Panel, but I do have MSI Afterburner so could do it from there if that would be helpful.

It's such an easy thing to do and also to reverse if it doesn't work (even more so if you already have the software to do it) that I don't see why you shouldn't give it a shot.