Ask GN 17: Why Only 99% GPU Utilization? (& Coil Whine)
Ask GN 17: Why Only 99% GPU Utilization? (& Coil Whine)
2016-05-11
everybody we're back for another ask
Jian episode I am freshly back from
Austin Texas where we did that and video
preview of the GT X 1080 you can check
all that content on the channel already
but this is our series where we answer
some of the questions that you all
submit so if you have a question post it
below I will try to get to it I will say
last time there were some very good
questions I'm not gonna get to them all
today if I miss it you want to try again
leave it below and I'll see if we can do
in the next video
today he starts off with a weird name
its ex ex Sharik xx YY oh I guess he
says or she says hey Steve went under
full load why is the GPU maxed out at
only 99 percent when the CPU can be at
100% so I actually checked on this
because it's a really good question I
didn't know the answer to and I emailed
that some of the people at Nvidia talked
to Jo vivoli who said that it basically
comes out to a rounding error so that's
all there is to it it's effectively 100%
load and by gpu-z
will report that as 99 percent loads
just a rounding thing but you can
basically imagine it as 100% so that's
the quick answer to that one the next
question is from exiled storm
who says quick question that you could
probably address a couple of seconds
unlikely I have an MSI z97 gaming 5
motherboard and I noticed when in the
EOC settings on the BIOS that there is a
memory try it mode and I'm not sure what
it does or what it means what is it so
memory try it is MSI's way of really
quickly and sort of lazily tightly
timing the memory timings so there's the
cache latency CAS latency and if you
have maybe 11 13 13 latency on your
memory you can do memory try it for cats
11 and it will just quickly tie in all
of those tertiary timings and all the
low-level times that are harder to work
with manually and and basically take
care of it for you so that's all that is
you would basically use it if you're
trying to do some sort of lazy mode
memory overclocking there's nothing
wrong with that but that's that's all
that's for next question amir mahmoud
who says i have a question for the next
ask GN i would find an answer to any
where all caps and it slowly become the
bane of my existence my pcs fans are
connected to the motherboard via three
pin fan
and controlled via BIOS so this all done
through BIOS with a user made custom fan
curve but that's that's not really
controlled by BIOS with a user made
custom fan curve the fans will randomly
ramp up to 100% and spin at that speed
for a few seconds and then slow down and
then ramp back up the only fix I've
found is to increase the minimum fan
speed in BIOS slowly so it sounds like
this boils down to something with the
fans I don't know what motherboard
you're using but some fans will have
sort of a minimum speed where the motor
can sustain the rotation and if you fall
below that speed then it's it's just not
gonna work you'll either depend on the
fan either turn itself off or it'll spin
at a higher speed than you asked for
because it can't sustain certain speeds
at a certain rpm and depending on what
the motor is capable of pushing so that
might be it might just be a case of
needing better fans for pin pwm header
might help you out a bit the three pin
headers that are controlled through the
motherboard are basically just being put
on to a 3 volt 5 volt or a 12 volt rail
from the power supply going into the
board so if your fan speed doesn't
happen to really fit on one of those
very well then you might experience all
the issues you're talking about
otherwise it could be something with the
motherboard
maybe a BIOS update well that's where I
would start as those two things next
question is demagogue who says a gamer
is Nexus can you touch on how to buy
quality products that do not produce
coil whine this is a big problem for me
when I'm bowing for power supplies and
stuff like that there's another part of
this question too for the first part
coil whine is an issue especially with
diffuse where if you're producing maybe
3000 FPS on a menu screen for example in
The Witcher a lot of cards will start
whining pretty high-pitched noise and
that's just from the coils in the card
being unhappy as far as looking for
something that doesn't produce that it
really just comes down to reading
professional reviews so for power
supplies we don't review them because to
do it properly I would want to buy like
10 thousand dollars of testing equipment
so we don't review those the people I
trust he do review them would basically
just be Johnny guru I would check Johnny
gurus reviews on a power supply
you buy it if you're really concerned
about coil wine and see if he was able
to produce some and if not then you're
probably good shape but that's really
all it comes down to there there's no
secret trick to what components do these
manufacturers use it's it's really just
how does it perform in the review and
what do people think of it and in our
video cards you card reviews will note
that if it has coil wine or something as
well
quality products you're basically
looking for things like solid capacitors
on them some people call them black
capacitors some people call them
Japanese capacitors but you want solid
capacitors higher-end vrm or electrical
design and that's stuff that really
without looking at it and knowing what
you're looking at you you wouldn't be
able to gauge what's better than the
other but manufacturers of course will
all say that they have the best
electrical design so you can't really
just take their word for it either
so it just boils down to reviews now the
second part of your question do you
think some companies are lying to us
about stable overclocked speeds at 1400
megahertz I guess where it's actually
not stable for the components and where
I should have gone with a 980 I at 1,100
clock versus the overclocked one I so
there's a couple things here over clocks
depend on what's called the silicon
lottery so there's sort of a limitation
on every GPU and everyone is different
or every CPU they all have different
voltage allowances and frequency
tolerances and it depends on the ASIC
quality depends on just how well your
chip Bend because they're not created
equal and that's not an intentional
thing it's just that it's silicon
manufacturing isn't a perfect process so
that's certainly some of it they may
have if your case is personal it may be
that they were able to work on a higher
or a better bin chip than you got in
your retail card whether or not I would
use that I don't know that I would use
the phrase line to us when you're
talking about over clocks most
manufacturers should be pretty
conservative if they're listing a number
that they think is achievable you I know
Kingston and Corsair are both fairly
conservative in their memory over clocks
for example where they'll say what's
realistically expected for that
particular chip as for video cards 14100
is
megahertz for certain devices like a
980ti is certainly sustainable we've
been able to sustain it but it does
depend on what you put for voltage
depends on the vrm design of the card
not all cards have the same vrm some
have more phases depends on the chip
itself silicon lottery again so if
there's a lot of stuff at play but I
don't I wouldn't go out and say that
companies are intentionally lying to you
it just it there's a lot of stuff moving
around but you can check our reviews we
do overclock and we'll list like what
our stable clock rate was after efforts
to do a high OC and that should inform
your decisions about 1400 I will say
1400 1450 megahertz is fairly high to
sustain in the long term like a couple
of years but if you're just going for an
immediate short-term overclock maybe a
year or so of uses usage then you can
get away with something like that on
most devices I've personally tested last
question for this week is from ill to
Xbox who says hey Steve I have a
question on my since I was thinking of
upgrading from my cheap standard Mouse
to something better suited for gaming
how much does polling rate actually
matter is 500 Hertz good enough or will
I actually benefit from 1000 Hertz and
games is 1,000 Hertz of meaningless
marketing strategy thinking of buying
the g6 r2 whereas the similarly priced
yrg 502 has a 1,000 Hertz polling rate
the g6 co2 has over 500 spline rate so
we talked about this a lot in the G 900
review that we recently posted first of
all I would suggest checking that out
and that's for everyone because that has
some pretty cool engineering stuff we
talked about that we don't normally go
into but with regard to this question
specifically I am NOT the best person to
purr to answer if 500 Hertz raises a
thousand Hertz actually really creates a
perceivable difference because I'm not
I'm not at that level of competitive
twitch gaming where I would see it
myself I would be better off improving
my own skill first but if you're a
really competitive gamer then maybe I
will tell you that the difference is
effectively one millisecond so if you
have a 1000 Hertz polling rate it means
that your mouse is communicating that
input and all of the movement to the
computer every one millisecond and
that's different from from the click
latency and things like that this is
specifically talking about the latency
from the the device the mouse to the
system or to the host so a thousand
Hertz is one millisecond pull a response
time 500 Hertz it's going to be 2
milliseconds half that speed so there's
there's a one millisecond difference
there whether or not you can see that
top to say I will tell you that I
personally do not see it I just not I
don't know maybe you have to be like a
pro eSports player or something but even
when I was very competitive with
Starcraft 2 I didn't see that difference
really so that's my take on it certainly
if people who are watching this think
that they can tell the difference leave
a comment below maybe help ill to Xbox
out with what you do and how you notice
it as far as the next part the g6o
tubers the G what is it G 502 I really
like both the g6 r2i I haven't used in a
while but it was one of my favorite mice
when I did use it it's got a good feel
to it it has a good weight it's not
nearly as light as the G 900 if I
remember correctly the G nine hundreds
pretty impressive but it's also $150 so
that would certainly be the best of both
worlds if you want to drop 150 bucks on
a mouse
but for most people I think you'd be
happy with G 602 or g502 if you're
looking strictly at Logitech if you want
to look outward of course there's m65
Pro we're working on reviewing right now
it's another one to consider or cheaper
versions Logitech G 303 something like
that SteelSeries sensei but yeah I like
G 602 just fine so it just depends on on
what you're going for I guess but for
things without a wire and that aren't
$150 I would definitely consider it
strongly that's all for this week I
don't really have any hair or beard
questions from you all I'm very
disappointed as you can see I grew one
over the past few days anticipating such
questions
but alas you've left me to have your
comments below if you have questions for
computer hardware or whatever else
thanks for watching pay traveling
special video I'll see you all next time
We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.