WEBVTT

00:00:02.159 --> 00:00:08.280
only one motherboard because your system

00:00:05.759 --> 00:00:12.660
needs only one CPU but with so many options all with different

00:00:10.139 --> 00:00:17.940
configurations of power phases PCI Express Lanes and m.2 Slots how do you

00:00:15.480 --> 00:00:22.080
choose which one is truly the best is it actually worth it to spend more money on

00:00:20.279 --> 00:00:26.039
a higher end motherboard well Micro Center sponsored this video and sent us

00:00:24.420 --> 00:00:31.199
10 motherboards from various manufacturers to test and um let me put

00:00:28.800 --> 00:00:36.239
it this way if the results surprised us then they're almost definitely going to

00:00:33.719 --> 00:00:41.700
surprise you like this was not the Biggest Loser not by a long

00:00:40.680 --> 00:00:45.480
shot and neither is Micro Center get the best

00:00:43.800 --> 00:00:48.719
prices and best selection on PC hardware and other technology at any of

00:00:46.860 --> 00:00:53.100
microcenter's 25 locations across the United States new customers can get a

00:00:50.700 --> 00:00:57.920
free 240 Gigabyte SSD at Micro Center offer valid in-store only no purchase

00:00:55.440 --> 00:00:57.920
necessary

00:01:04.040 --> 00:01:11.040
for every motherboard we tore down and rebuilt the exact same test bench we're

00:01:08.460 --> 00:01:15.600
using the same memory CPU Cooler power supply SSD and graphics card we even

00:01:13.439 --> 00:01:20.340
made sure to grab the exact same type of thermal paste for the last few boards

00:01:17.760 --> 00:01:24.119
when our first tube ran out this is to control our variables as tightly as

00:01:21.960 --> 00:01:27.659
possible we even tested every system in the same climate controlled environment

00:01:25.680 --> 00:01:30.840
after building each bench we'll have affiliate links down below for

00:01:28.860 --> 00:01:34.740
everything we use by the way we immediately found an important reason to

00:01:33.119 --> 00:01:39.360
choose one board over the other not every UEFI BIOS is created equally some

00:01:37.979 --> 00:01:44.820
of these didn't even allow us to change the pl1 and pl2 power limits that much

00:01:41.640 --> 00:01:46.439
if at all and while most of them did we

00:01:44.820 --> 00:01:50.520
had to dig through advanced CPU settings to find them in some cases after

00:01:48.180 --> 00:01:54.360
powering on we updated the BIOS enabled XMP and maxed out our long duration

00:01:52.380 --> 00:01:57.840
power limits whenever possible we could have gone for a more you know

00:01:56.220 --> 00:02:02.040
vanilla configuration here but given this is a more Enthusiast oriented

00:01:59.399 --> 00:02:06.479
exploration and that we are looking for benefits of more premium motherboards we

00:02:04.680 --> 00:02:12.420
felt this was the right approach the reality is modern CPUs have gotten

00:02:09.840 --> 00:02:16.319
really good at knowing exactly when it is okay to boost and when it is not

00:02:14.099 --> 00:02:20.340
these days it's actually pretty hard to break a CPU and on most boards you'll

00:02:18.599 --> 00:02:24.060
get some kind of warning when you toy with the settings that can do so

00:02:21.720 --> 00:02:28.200
Enthusiast cooling has also gotten really good with most performance tier

00:02:25.860 --> 00:02:33.540
gaming systems being equipped with a 240 millimeter AIO water cooler or a large

00:02:30.780 --> 00:02:36.780
Tower heatsink we used a Noctua nhd15 which will run you about 115 bucks at

00:02:35.400 --> 00:02:40.680
Microcenter and should give our processor plenty of room to stretch its

00:02:38.640 --> 00:02:44.760
legs while providing some incidental cooling to the vrms which could help

00:02:42.900 --> 00:02:49.560
performance something to note is that even with a great cooler our cinebench

00:02:47.099 --> 00:02:53.519
and blender tests never saw our CPU push past the 250 watt Mark and this is in

00:02:52.019 --> 00:02:59.040
spite of setting the power limits to the maximum like 4 thousand watts in some

00:02:55.680 --> 00:03:01.379
cases so what gives with that well the

00:02:59.040 --> 00:03:05.400
thing is even if we do find some performance differences from board to

00:03:02.819 --> 00:03:09.360
board changing the power limits is kind of like removing the governor on a car

00:03:06.800 --> 00:03:13.860
yes theoretically the speedometer goes up to 300 miles an hour but your engine

00:03:11.819 --> 00:03:17.700
is still going to max out on its own and like occur you can open up some more

00:03:15.599 --> 00:03:21.599
Headroom by increasing voltage adjusting load line calibration and manually

00:03:19.440 --> 00:03:26.640
changing the frequency of your chip but that's topic for another video like this

00:03:23.940 --> 00:03:30.900
one or for you to explore by subbing to an extreme or clocker like your Bauer

00:03:28.379 --> 00:03:34.319
this video is all about out of the factory performance and right out of the

00:03:33.060 --> 00:03:39.599
gate I'm really surprised our big winner was

00:03:36.780 --> 00:03:43.319
MSI's gone like z590 which ended up with the highest cinebench score and the

00:03:41.340 --> 00:03:48.720
fastest blender render times to be clear this is a 900 motherboard competing

00:03:46.980 --> 00:03:52.200
against boards in the 100 to 200 category but I mean even in the days

00:03:50.819 --> 00:03:55.920
when motherboard's affected system performance by a lot it was common to

00:03:54.540 --> 00:04:00.299
see expensive boards that were like loaded up with features in this case

00:03:57.599 --> 00:04:04.080
Thunderbolt 4 Wi-Fi 6E obviously super Premium Power delivery but it didn't

00:04:02.459 --> 00:04:07.920
actually offer a clear performance Advantage unless you were trying to

00:04:05.459 --> 00:04:11.879
break Sub-Zero overclocking records of course that performance Advantage was

00:04:09.420 --> 00:04:15.120
still pretty small so let's take a look at the board that came in second place

00:04:13.260 --> 00:04:20.760
at least in terms of cinebench our ASRock z590 Pro 4 was behind by just 19

00:04:18.600 --> 00:04:25.139
points in cinebench and ran only a couple of seconds slower in the renders

00:04:22.199 --> 00:04:29.940
yet managed a whopping one percent more FPS in CS go okay that that last bit

00:04:27.840 --> 00:04:36.240
doesn't sound that impressive but it costs a fraction as much at 185 US

00:04:32.759 --> 00:04:39.000
Dollars and while it only has 14 power

00:04:36.240 --> 00:04:43.440
phases with 50 amp chokes 2.5 gigabit Lan and no onboard Wi-Fi the choice is

00:04:41.520 --> 00:04:48.419
pretty clear if you care about value buy one of these instead and maybe use the

00:04:46.320 --> 00:04:51.479
money you saved at ltstore.com the beanie and the water bottle are a great

00:04:49.680 --> 00:04:55.919
combo as for the rest of our results they're mostly what we were expecting

00:04:53.520 --> 00:05:00.120
this CPUs are more integrated than ever these days so motherboards don't

00:04:57.479 --> 00:05:04.979
contribute like they used to but there are some outliers for instance MSI wins

00:05:02.880 --> 00:05:09.360
the crown for the best microatx b560 board as well beating out three other

00:05:06.960 --> 00:05:13.740
z590 boards at least in cinebench the b560m mag mortar was slower to render by

00:05:11.880 --> 00:05:18.479
eight or seven percent in BMW and classroom respectively but it had the

00:05:16.560 --> 00:05:25.380
third highest cinebench score the second highest FPS in CS go and the second

00:05:21.540 --> 00:05:27.600
fastest turn time in Civ 6. b550s aren't

00:05:25.380 --> 00:05:31.220
supposed to overclock so picking up a supposedly lower tier board for its

00:05:29.580 --> 00:05:39.960
performance is a little unintuitive meanwhile MSI z590a Pro lost instead of

00:05:35.340 --> 00:05:41.340
inch by just 17 points but how well you

00:05:39.960 --> 00:05:45.660
could explain this with some simple run-to-run variants that's such a narrow

00:05:43.500 --> 00:05:51.060
Victory that's barely one at all you'd think that but we actually re-ran our

00:05:48.479 --> 00:05:55.080
benches on every board and these margins between them while narrow were

00:05:53.039 --> 00:05:58.860
surprisingly consistent one of the major benefits we saw with MSI boards was when

00:05:57.120 --> 00:06:03.120
you set them up it asks you what kind of cooling solution you have the more

00:06:00.900 --> 00:06:07.500
robust solution you pick the higher they push both the long and short duration

00:06:04.979 --> 00:06:10.919
power limits since an nhd-15 is basically as good as a solid water

00:06:08.820 --> 00:06:15.000
cooling setup we chose water cooled instead of tower cooler and it maxed out

00:06:12.780 --> 00:06:18.600
the limits for us at 4 000 Watts you could easily achieve the same on a

00:06:16.440 --> 00:06:22.259
different brand but this kind of user friendliness is something the others

00:06:19.919 --> 00:06:27.479
could learn from of course performance isn't the entire story MSI's z590 board

00:06:25.620 --> 00:06:33.120
managed to stay about 10 degrees cooler on the vrms so while the 90 plus degrees

00:06:30.300 --> 00:06:38.520
of our b560m mag mortar isn't out of spec for these kinds of components a hot

00:06:35.280 --> 00:06:40.740
day could easily push that to 100 or 110

00:06:38.520 --> 00:06:44.819
and remember we have some airflow around the socket if you're using an AIO or

00:06:42.419 --> 00:06:48.780
you're in a tight case you might not as for our losers

00:06:46.440 --> 00:06:54.780
well as you could have guessed it's mostly the b560s and our solitary h510

00:06:51.840 --> 00:06:58.139
board the Gigabyte z590 UD was 20 points behind at cinebench and so it

00:06:56.520 --> 00:07:01.919
unfortunately gets to hang out down here despite performing significantly better

00:06:59.699 --> 00:07:06.900
in the render times and perform the best when it comes to CS go at nearly 560 FPS

00:07:04.740 --> 00:07:10.800
average so if you're purely after gaming performance this one might call to you

00:07:09.060 --> 00:07:14.400
from Micro Center shelf or in one of their custom prebills despite being down

00:07:12.300 --> 00:07:19.319
here in the losers bracket as for the one that was bad across the

00:07:16.860 --> 00:07:23.580
board there are a few reasons for this first off is their inability to set

00:07:21.300 --> 00:07:28.380
higher power limits most of them capped out at 125 150 or 200 watts and

00:07:27.000 --> 00:07:32.160
increasing the duration limit didn't help all that much either most of them

00:07:30.180 --> 00:07:37.800
aren't running too hot and if you chose a matching mid-tier CPU they would most

00:07:35.220 --> 00:07:42.780
likely be fine but they're just not able to live up to the 11 900k's full 250

00:07:40.560 --> 00:07:46.319
watt potential they typically have fewer power phases and a less robust power

00:07:44.400 --> 00:07:50.520
delivery system with just one eight pin EPS connector speaking of phases we're

00:07:48.479 --> 00:07:55.259
building out a DIY phase change CPU Cooler with some help from Brian the

00:07:53.220 --> 00:07:59.639
electrician so make sure you get subscribed so you don't miss it back to

00:07:57.060 --> 00:08:04.620
our b560 boards though these power draw limits aren't surprising but what is

00:08:02.099 --> 00:08:11.220
surprising are the bottom two boards you might be shocked to learn the ASRock

00:08:07.400 --> 00:08:13.440
h510m wasn't in last place it has no XMP

00:08:11.220 --> 00:08:18.319
option and memory can only go as high as 3200 megahertz so it did see the worst

00:08:15.599 --> 00:08:26.639
CS go score at 407 22 FPS average however it absolutely crushed our loser

00:08:23.160 --> 00:08:29.280
board in cinebench by roughly 4

00:08:26.639 --> 00:08:33.300
000 points render times were also won by nearly an extra minute in the BMW render

00:08:31.379 --> 00:08:37.800
and two minutes in the classroom as impressive as this is these results are

00:08:35.399 --> 00:08:43.260
still significantly lower than the b560m board above it but what the heck is

00:08:40.620 --> 00:08:47.519
going on with our Gigabyte b560m ds3h well power phases don't seem to matter

00:08:46.140 --> 00:08:52.200
all that much until they do this thing starts out

00:08:49.920 --> 00:08:57.540
strong and the initial render in cinebench R23 finishes quickly giving

00:08:55.019 --> 00:09:03.380
you hope but with each subsequent render the score goes down and down and down

00:09:01.140 --> 00:09:07.500
all the way to about 7688 points

00:09:06.420 --> 00:09:11.519
ouch if we check our clock speeds in power

00:09:09.600 --> 00:09:14.519
they're all over the place while CPU temperatures are mediocre at 80 to 84

00:09:13.500 --> 00:09:22.920
degrees what gives well our vrm temperatures are terrible

00:09:19.560 --> 00:09:25.860
quickly hitting 112 degrees Celsius this

00:09:22.920 --> 00:09:31.140
board is trying to squeeze all the power it can out for our CPU but the load is

00:09:29.459 --> 00:09:34.980
simply too much for the system to bear and it buckles onto the pressure

00:09:32.480 --> 00:09:37.800
dropping wattage and clock speeds in an effort to lower the temperatures

00:09:36.120 --> 00:09:42.240
bringing them back up to see if it's okay yet and then quickly settling back

00:09:39.779 --> 00:09:46.440
down again it's like a person trying to take a bath in boiling water the toe

00:09:44.040 --> 00:09:50.160
dips in they recoil and wait a moment and then they they try again and it just

00:09:48.420 --> 00:09:53.820
keeps going like that our other CPU heavy tasks saw similarly poor

00:09:51.959 --> 00:09:59.160
performance with this board seen the worst turn times in Civ 6 and the second

00:09:55.980 --> 00:10:00.899
worst FPS in CS go I have a hunch that

00:09:59.160 --> 00:10:05.700
if the memory had been set to 3200 megahertz it would have lost the battle

00:10:02.940 --> 00:10:10.620
to the ASRock h510m as well so our winner is obvious but it's not the

00:10:08.279 --> 00:10:15.720
Godlike as it stands the best board of the bunch with prices we can actually

00:10:12.060 --> 00:10:17.459
recommend is the ASRock z590 Pro 4. it

00:10:15.720 --> 00:10:22.440
didn't come out on top in any category but it was always near the head of the

00:10:19.920 --> 00:10:26.519
pack at 185 dollars the BIOS was relatively easy to use it

00:10:24.720 --> 00:10:30.600
has a lot of options for expansion and storage and it comes with asrock's base

00:10:28.920 --> 00:10:34.620
frequency boost technology that supposedly pushes non-k CPUs past their

00:10:32.700 --> 00:10:37.680
limit through base clock overclocking so theoretically you should be able to buy

00:10:36.060 --> 00:10:40.800
this board put a lower end chip in it and still get pretty good performance so

00:10:39.240 --> 00:10:45.839
you can afford something better our honorable mention is MSI's b560m mag

00:10:43.560 --> 00:10:50.220
mortar for a smaller machine on a budget it's a great choice just like Micro

00:10:48.120 --> 00:10:53.279
Center today's sponsor get the best prices and the best selection on PC

00:10:51.839 --> 00:10:57.899
hardware and Technology at any of Micro Center's 25 locations across the United

00:10:55.079 --> 00:11:01.560
States new customers can get a free 240 Gigabyte SSD at Micro Center offer valid

00:10:59.940 --> 00:11:05.040
in-store only no purchase necessary thanks for watching guys if you enjoyed

00:11:03.240 --> 00:11:08.579
this video make sure to check out our how motherboards work turbo nude Edition

00:11:06.779 --> 00:11:12.380
for a better explanation of how some of this works I'll have the link down below