Do I have to upgrade?

[mrtwonavels]

I have a Msi pro h610mg ddr4 mobo and was wondering if it would be worth it to upgrade or not? Would I get a performance boost if I get a higher Teri mobo or should I just instead save up and buy a 3060

[Middcore]

Unless your CPU is being gimped by poor VRM (and you don't mention your CPU at all), buying a better motherboard will not by itself make any difference in performance.

[mrtwonavels]

1 minute ago, Middcore said:

Unless your CPU is being gimped by poor VRM (and you don't mention your CPU at all), buying a better motherboard will not by itself make any difference in performance.

What do you mean don’t mention your cpu?

[Agall]

Just now, mrtwonavels said:

What do you mean don’t mention your cpu?

Depends on the CPU, if you're running a 12100f or something nearby, then no. If you've got a 13900k in it for some reason, then yes.

[mrtwonavels]

1 minute ago, Agall said:

Depends on the CPU, if you're running a 12100f or something nearby, then no. If you've got a 13900k in it for some reason, then yes.

Alright I got a 12600k

also whe I open ups task manager during games it says 98 percent usage is this a problem that my power supply is weak or just that its a 12th gen i5

(I have a 500watt power supply)

[ThePCNerd129]

11 minutes ago, mrtwonavels said:

I have a Msi pro h610mg ddr4 mobo and was wondering if it would be worth it to upgrade or not? Would I get a performance boost if I get a higher Teri mobo or should I just instead save up and buy a 3060

I have a an rtx 3060... Things great just save for a 3060

[mrtwonavels]

Just now, ThePCNerd129 said:

I have a an rtx 3060... Things great just save for a 3060

Will do

[GuiltySpark_]

11 minutes ago, mrtwonavels said:

also whe I open ups task manager during games it says 98 percent usage

What games are you seeing 98% usage on a 12600k?

[Agall]

16 minutes ago, mrtwonavels said:

Alright I got a 12600k

also whe I open ups task manager during games it says 98 percent usage is this a problem that my power supply is weak or just that its a 12th gen i5

(I have a 500watt power supply)

You're probably holding back that 12600k with a low end board like that. It might actually be worth looking at a Z690 motherboard which have gotten pretty cheap recently. That board supposedly has a 110W hard limit on the CPU power draw.

 

Minimum suggested for an RTX 3060 is 450W, so a quality 500W PSU should be fine. At that wattage, it would be worth telling us what specific 500W PSU you have to make a proper suggestion on if that's worth replacing.

[PDifolco]

13 minutes ago, mrtwonavels said:

Alright I got a 12600k

also whe I open ups task manager during games it says 98 percent usage is this a problem that my power supply is weak or just that its a 12th gen i5

(I have a 500watt power supply)

12600K on a H610 board is a weird choice, I don't think the board can handle the CPU at full throttle, and you can't even profit of any OC...

[Agall]

1 minute ago, PDifolco said:

12600K on a H610 board is a weird choice, I don't think the board can handle the CPU at full throttle, and you can't even profit of any OC...

The internet says that the board has a 110W TDP limit, so being at 100% usage in gaming at a lock 110W TDP isn't far fetched.

[mrtwonavels]

3 minutes ago, Agall said:

You're probably holding back that 12600k with a low end board like that. It might actually be worth looking at a Z690 motherboard which have gotten pretty cheap recently. That board supposedly has a 110W hard limit on the CPU power draw.

 

Minimum suggested for an RTX 3060 is 450W, so a quality 500W PSU should be fine. At that wattage, it would be worth telling us what specific 500W PSU you have to make a proper suggestion on if that's worth replacing.

Evga 500 gd

[PDifolco]

4 minutes ago, Agall said:

The internet says that the board has a 110W TDP limit, so being at 100% usage in gaming at a lock 110W TDP isn't far fetched.

Yes but some H610 perform like crap and throttle CPUs - not sure for a 12600K tho, it's not that powerful, but imo that CPU merits at least a B660

[Agall]

21 minutes ago, mrtwonavels said:

Evga 500 gd

You're fine on that front, its actually a 500W PSU. Sometimes that's not the case.

 

B660 won't give you overclocking, Z690 boards are inexpensive now and will allow you to overclock that 12600k if you wanted to.

[mrtwonavels]

Great so I just buy a z60 and 3060

[Tetras]

43 minutes ago, Agall said:

The internet says that the board has a 110W TDP limit, so being at 100% usage in gaming at a lock 110W TDP isn't far fetched.

I'd guess a 12600K shouldn't throttle much (if at all) with a 3060.

 

In TPU's 7800X3D review the average gaming power draw (of the 12600K) was 59 watts. I see more than that in YouTube videos, generally around 60-80 depending on the game/settings, but it rarely gets to as high as 90.

[Agall]

Just now, Tetras said:

I'd guess a 12600K shouldn't throttle much (if at all) with a 3060.

 

In TPU's 7800X3D review the average gaming power draw was 59 watts. I see more than that in YouTube videos, generally around 60-80 depending on the game/settings, but it rarely gets to as high as 90.

7800x3D power draw isn't comparable to a 12600k.

 

In the limiting scenarios where it'll want to boost to its max boost clock, it'll go up to 200W. Cutting that maximum wattage in about half will chop a solid 20-30% performance off the top.

 

The choice in GPU doesn't matter in plenty of games, plenty of which see 60 fps caps due to CPU limitations with even the top end CPUs.

[Tetras]

2 minutes ago, Agall said:

7800x3D power draw isn't comparable to a 12600k.

I know. They include the average games and applications power draw of all their recent test CPUs in the review, whereas the original 12600K review didn't include these results.

 

2 minutes ago, Agall said:

In the limiting scenarios where it'll want to boost to its max boost clock, it'll go up to 200W. Cutting that maximum wattage in about half will chop a solid 20-30% performance off the top.

I don't think that is likely to happen often, if at all, when gaming, though the OP hasn't stated their usage so I'm just assuming it is gaming.

 

3 minutes ago, Agall said:

The choice in GPU doesn't matter in plenty of games, plenty of which see 60 fps caps due to CPU limitations with even the top end CPUs.

It does, because the 12600K is less likely to be maxed out in high CPU demand scenarios with a GPU like that, whereas e.g. with a 4090 playing esports at 1080p, that's more likely to be an issue and in the videos I watched (not that many, but eh) it is the only one (4090 pairing) I saw the CPU manage to exceed 100 watts:

 

[Agall]

8 minutes ago, Tetras said:

I know. They include the average games and applications power draw of all their recent test CPUs in the review, whereas the original 12600K review didn't include these results.

 

I don't think that is likely to happen often, if at all, when gaming, though the OP hasn't stated their usage so I'm just assuming it is gaming.

 

It does, because the 12600K is less likely to be maxed out in high CPU demand scenarios with a GPU like that, whereas e.g. with a 4090 playing esports at 1080p, that's more likely to be an issue and in the videos I watched (not that many, but eh) it is the only one (4090 pairing) I saw the CPU manage to exceed 100 watts:

 

There's a fallacy you seem to be prescribing to. Unless there's a hard GPU limit, there's going to be performance left on the table in most game. The more significant ones are those that are almost always at a hard CPU limit, like MMO/multiplayer type games with a ton of CPU draw calls that might have a 60 fps maximum in some scenarios with even a 7800x3D or 14900k.

 

I have to beat this horse to death, but in this case, there are games who's minimum framerates are dramatically worse with a lesser CPU, lesser being one with less IPC/frequency/cache. They're not multicore limited, but single core and memory bandwidth (which includes cache) limited. The best example I've seen being WoW.

[Tetras]

1 minute ago, Agall said:

There's a fallacy you seem to be prescribing to. Unless there's a hard GPU limit, there's going to be performance left on the table in most game. The more significant ones are those that are almost always at a hard CPU limit, like MMO/multiplayer type games with a ton of CPU draw calls that might have a 60 fps maximum in some scenarios with even a 7800x3D or 14900k.

 

I have to beat this horse to death, but in this case, there are games who's minimum framerates are dramatically worse with a lesser CPU, lesser being one with less IPC/frequency/cache. They're not multicore limited, but single core and memory bandwidth (which includes cache) limited. The best example I've seen being WoW.

I don't know what you're telling me I'm wrong about specifically.

 

I literally just watch a bunch of YouTube videos and then note their typical power consumption.

 

I'm only interested (in this context) in the gaming power draw and if the OP should/needs to buy a new motherboard to handle it.

 

For most of the videos their numbers are higher than TPU's review (@ 59 watts), but still manageable for a low-end board, I should think.

 

That one with a 4090, they're quite high (around 100 or higher than that, on average), but that's an outlier from the trend in the videos I watched.

 

Higher performing graphics card with a lower resolution, usually results in higher CPU power consumption, but I agree that certain games like MMOs (depending on their age), or maybe strategy/RTS are likely to have a higher power draw on average, even if the graphics card is not a high-end one.

[Agall]

2 minutes ago, Tetras said:

I don't know what you're telling me I'm wrong about specifically.

 

I literally just watch a bunch of YouTube videos and then note their typical power consumption.

 

I'm only interested (in this context) in the gaming power draw and if the OP should/needs to buy a new motherboard to handle it.

 

For most of the videos their numbers are higher than TPU's review (@ 59 watts), but still manageable for a low-end board, I should think.

 

That one with a 4090, they're quite high (around 100 or higher than that, on average), but that's an outlier from the trend in the videos I watched.

 

Higher performing graphics card with a lower resolution, usually results in higher CPU power consumption, but I agree that certain games like MMOs (depending on their age), or maybe strategy/RTS are likely to have a higher power draw on average, even if the graphics card is not a high-end one.

Steady state power consumption isn't the whole picture on a motherboard's capabilities with a CPU. 12600k will draw as much power as its allowed, in the case of a budget motherboard, about half of what it wants to draw. Either way, that's a loss in performance where the speculation is how much. If that prevents +500MHz boost clock due to UEFI/TDP limitations, then it can be substantial in the most limiting scenarios to framerate, like the massive dips you can find in hard CPU limiting scenarios.

 

There's a reason why people don't use games like MMOs/multiplayer games to test CPUs, the scenarios are hard to control and are usually variable based on the testing environment, to include the specific scenario being tested. That's where it requires a lot of knowledge, time, runs, and various locations to fully test those games, something that isn't worth the time of academic benchmarkers. So they stick to AAA single player games with less variability.

 

The chip can draw in excess of 300W, will try to draw in excess of 200W. Giving it an arbitrary TDP limitation as low as 110W will limit the performance in the scenarios where it wants to spike power, which are usually the scenarios where you're CPU limited. Its a great way to get noticeable framerate dips. Ive even seen it when doing CCD0 vs CCD1 testing in games that REALLY love 3D v-cache, the experience goes from buttery smooth to slight spiking, considering that the 7950x3D is 150W TDP limited, even though CCD1 on my CPU can boost as high as 5.9GHz, it can't draw the 200W it wants to in that scenario. Something 3D v-cache doesn't have to do since it can compensate with 3x the L3 cache.

[Tetras]

1 minute ago, Agall said:

[...snip]

Aha, so what you're saying is: the loss of performance due to curtailing the turbo/power spike could result in significant frame drops in CPU-limited scenarios that aren't necessarily visible to conventional reviews/testing. And there are some games in particular, that it is more important.

 

Still, my edumacated guess from watching the YouTube videos, is that I wouldn't bother upgrading the motherboard, as I think if 110 watts is available, that's going to be sufficient in most cases.

 

A gamer that mainly plays the type of games you mention, is more sensitive to stutters and/or has a high-end video card and/or plays at low resolutions with high average fps are more likely to care about the drops, I'd imagine.

[Agall]

24 minutes ago, Tetras said:

Aha, so what you're saying is: the loss of performance due to curtailing the turbo/power spike could result in significant frame drops in CPU-limited scenarios that aren't necessarily visible to conventional reviews/testing. And there are some games in particular, that it is more important.

 

Still, my edumacated guess from watching the YouTube videos, is that I wouldn't bother upgrading the motherboard, as I think if 110 watts is available, that's going to be sufficient in most cases.

 

A gamer that mainly plays the type of games you mention, is more sensitive to stutters and/or has a high-end video card and/or plays at low resolutions with high average fps are more likely to care about the drops, I'd imagine.

There's an 'ignorance is bliss' argument here. An example of where I tested this myself was in 7950x3D vs 4790k RTX 4090 OC'd 4K ultra Warframe testing I did. The stutters I experienced likely wouldn't be as noticeable by a user who isn't used to literally the best PC gaming experience.

 

 

This game is highly optimized, where I was even able to later get the game to run at 4K low on my main rig at over 1400 fps in the same scenario I was able to reach 1200 fps at 1080p ultra. Overall, the 4790k did better than I expected, but there were noticeable limitations. This game does use a fair amount of CPU power, where I'll see 60% utilization on CCD0 in the most limiting scenario.

 

A 12600k can obviously reach higher than a 4790k, but compared to CCD1 testing, the minimum framerates of having effectively an R7 7700 TDP capped at ~120W (for a single CCD) was limiting. Its also thermally limited by CCD0 even if its turned off since it'll soak the IHS otherwise, so somewhat comparable TDP limitation that a 110W capped 12600k would be in.

 

The 7950x I have at work running a server is TDP capped to 105W, which did chop down a lot of its all core performance. At least with that server, its barely getting above 10% even under load, so a couple cores operating at 5.5GHz aren't going to get it to draw 120W almost ever (the application only uses one core per user and is single threaded bound, at most has 2-3 users but is licensed up to 16 users, so 7950x was a perfect fit. Doubled the performance going from a relatively new Xeon server). 

 

 

TLDR: The times it needs to spike past a TDP limit are where you're most limiting, especially if its half of desired draw. Some people may not care and are willing to sacrifice performance consistency for it, but I personally don't think its worth it. In OP's scenario, a Z series chipset would also allow for overclocking, which is a whole experience on its own and is worth it in my opinion.

[Tetras]

2 minutes ago, Agall said:

TLDR: The times it needs to spike past a TDP limit are where you're most limiting, especially if its half of desired draw. Some people may not care and are willing to sacrifice performance consistency for it, but I personally don't think its worth it.

 

9 minutes ago, Agall said:

The stutters I experienced likely wouldn't be as noticeable by a user who isn't used to literally the best PC gaming experience.

Yes, I think I see what you're saying and for a demanding user, curtailing their CPUs power draw is not a good move as a power/turbo limited board could have major implications for their overall gaming experience.

 

It is interesting and I have made a note of this thread, because when someone reports frame drop issues it is useful information to consider, but I think in this case, what I'm trying to do from watching the YouTube videos is to condense the likely operating power envelope of the CPU when gaming and translate that into the likelihood of those spikes causing unacceptable frame drops for the OP's PC.

 

From what I can see in the bulk of the videos, the average gaming power draw is pretty modest (though as you say, they tend to include only a few types of game in these videos) and I don't think it would need more performance very often and if it does, my guess (and I judge that partly from the hardware used, or likely to be used, which might be a mistaken assumption, depending on the OP) is that it is not significant enough to merit an upgrade to the motherboard.

 

If the OP were to upgrade to a high refresh monitor and a 4080 Super, for example, then my assessment is more likely to be incorrect.

[Agall]

10 minutes ago, Tetras said:

 

Yes, I think I see what you're saying and for a demanding user, curtailing their CPUs power draw is not a good move as a power/turbo limited board could have major implications for their overall gaming experience.

 

It is interesting and I have made a note of this thread, because when someone reports frame drop issues it is useful information to consider, but I think in this case, what I'm trying to do from watching the YouTube videos is to condense the likely operating power envelope of the CPU when gaming and translate that into the likelihood of those spikes causing unacceptable frame drops for the OP's PC.

 

From what I can see in the bulk of the videos, the average gaming power draw is pretty modest (though as you say, they tend to include only a few types of game in these videos) and I don't think it would need more performance very often and if it does, my guess (and I judge that partly from the hardware used, or likely to be used, which might be a mistaken assumption, depending on the OP) is that it is not significant enough to merit an upgrade to the motherboard.

 

If the OP were to upgrade to a high refresh monitor and a 4080 Super, for example, then my assessment is more likely to be incorrect.

I see about 40-45W on my 7950x3D in a game like Diablo 4, which uses a decent amount of CPU power. Its only about 3W extra to have CCD1 on standby in comparison to disabling it in the UEFI.

 

It really depends on the games more than the hardware, but most games can be generalized based on those academic style benchmarks. Where I come in and reignite my crusade is that most people don't spend most of their time playing those games, and the games they actually play (which are usually multiplayer games) are far more CPU limited than they realize. Being limited by IPC/cache/frequency and not multicore performance specifically, which can still cause spikes in power draw but steady state will have low draw.

 

The easiest example being WoW, where you could have a 1050ti and 7800x3D and likely still wind up in CPU limited scenarios. I can attest to a 5800x3D since that's the last CPU I owned when I still played that game that unlocked the minimum framerate in raids/battlegrounds/cities from like 40 fps with a 3950x to over 60 fps with a 5800x3D. The game in itself would go well above 200 fps in low population/NPC zones, which is where I was likely at a GPU limit (potentially) if not simply a higher CPU limited framerate (I know how to test this now, but I don't play the game anymore). Minding that was with a 6900 XT and 4K ultra as well.