NVIDIA Laptop GTX 1080 vs. 1070 Benchmarks [Initial]
NVIDIA Laptop GTX 1080 vs. 1070 Benchmarks [Initial]
2016-08-15
Notebook GPUs are finally crawling out
from under the M branding that's been on
nvidia and AMD GPUs for ages now and
that's because the architecture has
gotten to a point where it's efficient
enough that further changes to core
account clock rate things like that
don't need to be quite as Extreme as
they once were and so away goes the M
branded notebook GPU and that will be
happening for AMD shortly before getting
to that this coverage is brought to you
by our power and their new element
gaming PC which ships with a full
tempered glass side window arc LED fans
and LED underglow as seen in this video
today though we're talking about NVIDIA
and that's because they've got the new
gtx 1060 1070 and 1080 GPU is the same
GP 104 and 106 chips that we saw in the
desktop components now in notebooks and
that addition means that pascal almost
exactly identically to the desktop GPUs
will now be in notebooks we managed to
leverage about four hours of hands-on
demo time while at an Nvidia event so
got to test the new notebooks with our
normal benchmark suite that includes
frame time analysis and we did shadow of
mordor testing that was the only game
they really had their test on the 10 17
1080 I've got a 1060 notebook on the way
and then there's also a quick
overclocking note toward the end of this
video we're invited to give a hands-off
demo of overclocking but we'll see how
that works out with the release product
strictly comparing desktop and notebook
GPUs with the new Pascal architecture
the differences between the two are
limited this concept of sacking the M
line under nvidia branded GPUs started
with the GTX 980 non-m which we reviewed
in the Dominator Pro G last year and
it's continuing with Pascal the
architectures are the same the power
design of the GPU proper is the same the
cores are mostly the same memory has
been inspected time from the last gen of
mobile GPUs as well and is now at 8
gigabits per second for the 1070 and
1062 m gigabits per second with gddr5 X
on the 1080 notebook GPU there's one
core difference between desktop and
laptop GPUs and that's what the gtx 1070
the mobile gtx 1070 it has 2048 cores
rather than the 19
Montee chords on the desktop equivalent
and it's also reduced the clock rate to
1645 megahertz boosted from 1683 on the
desktop so more cores and a slightly
slower clock rate as for the rest the
GTX 1080 and 1060 have the same core
account as on desktop that's 2560 on the
1080 and 1280 on the 1060 and the GTX
1018 notebooks will boost at 17 33
megahertz stock the same as the desktop
and the 1060 boosts at 16 70 megahertz
stock all three GPUs have a range of
about 10% on the clocks for factory OC
by laptop manufacturers so these numbers
might be slightly higher in some systems
depending on if the notebook you're
looking at has been pre overclocked
there's no hard TDP number for
publication though Nvidia does have a
TDP spec for the manufacturers and the
reason there's no hard number is because
TDP as a reminder is more of a
measurement of the cooling requirements
for a component than the actual power
draw from the wall or elsewhere in the
system and for that reason because
manufacturers have different skin temp
specifications when you touch a laptop
they all have a different specification
on what's acceptable for heat they have
different thermal specifications and
requirements within their systems
there's no number for us to give you we
do know a little about the electrical
design though these use MXM cards as one
would expect for most of the larger GPUs
and notebooks and the MXM cards will use
similar power design to the desktop GPUs
including the same tooth power supply
and multi-phase power control in theory
this setup also increases power density
supporting higher GPU power in a smaller
space while further reducing total watt
draw from the battery during the voltage
and current conversion process we'll
explore this further when we have time
to take part one of the laptops into a
PCB analysis of the board and the MXM
cards there's also new display
technology getting pushed alongside
these GPUs including 120 Hertz 1440p
monitors and at least one of the
notebook lines coming out from the many
manufacturers supporting this lineup and
there's also a big push just towards 120
Hertz 1080p displays in notebooks or if
that's kind of on the high end and
there's a budget there's still the
normal 1080p 60 1080p 75 if you prefer
slightly faster refresh
and there's the normal flow of 4k
displays as well though that obviously
increases price pretty significantly
overclocking the laptop GPUs is done by
modifying core clock and memory clock as
it normal but make some changes one of
those is the elimination of power
percent target control or the ability to
increase the GPS power consumption at
the expense of higher thermals with
notebooks this is reasonably disabled
there's not much room in the thermal
envelope to play and we asked if
over-voltage and under-voltage would be
possible for notebooks receiving a
resolute No upon further discussion we
learned that the notebook components
will follow similar volt frequency
curves as discussed in our GTX 980 OC
review and that users may modify DPM
volt frequency configurations to some
extent but cannot exceed the maximum or
minimum values that are pre-configured
without over-and-under voltage will see
how useful manual volt frequency
plotting is thus far for enthusiasts we
haven't found a truly solid reason to
detour off the path of manual slider
overclocking for just the membrane core
we previously overclocked to the
Dominator Pro G laptop with some level
of success managing to achieve a 200
megahertz core offset on the GTX 980
non-m component as for what we were
shown in the hands off demo of the 10
series and video somewhat successfully
attempted to overclock a 20 63 megahertz
core on a gtx 1080 laptop with an offset
of 225 mega Hertz from the core the
demonstration didn't show memory
overclocking or any other changes fans B
control isn't yet available as far as we
understand but it will be through some
vendors and note that we also don't have
a good idea of how fan noise responded
to this overclock since the presentation
was in a noisy room we also only thought
on one game and there was some initial
software hurdles when trying to perform
the overclock so we'll validate all this
in our own reviews shortly and video
explicitly stated that their overclocked
laptop did not have a magic chip which
is something that they said also for the
gtx 980 launch we saw following that
launch that's pretty hard to hit the 21
14 megahertz advertised in the
presentation clock rate that was shown
on the original 1080 so we'll see if
that's true here as well but certainly
some of the founders Edition cards were
able to hit that clock rate so it's
within I suppose some level of normal
there
rants with the silicon lottery and just
other issues generally with thermals
things like that we don't know how fast
these fans are running don't know the
volume that they're running at just
because it's all contained within a
presentation environment now as for the
impact of the OSI during the hands off
demo dooms FPS range at 1080p ultra
loosely measured just from visual
observation of the FPS counter it seems
that dooms FPS went from 116 to 176 fps
when the 1607 megahertz stock clock rate
was in place and with the overclocked
220 63 megahertz core the game was
ranging from about 119 to 197 fps so we
saw visually just from loose observation
a gain of about 20 FPS or thereabout and
that's looking at min and Max values
without any measurement tools speaking
of measurement tools and video had about
a dozen laptop setup in their demo suite
and we were able to negotiate clearance
to run some of our own software for
benchmarking of the new laptops
including a 1070 and 1080 laptop a
cleveo unit and an msi unit and this
omitted the 1060 unfortunately but the
games for testing were not present on
the 1060 in the demo suite because we
used our own software and measurement
tools for the test we can trust the
results as accurate I know that there's
somewhat comparable to the testing we've
done at home base still only a few games
were present on the system and we ended
up testing shadow of mordor on a two
pass
setup due to time constraints and test
it across three normal resolutions 1080
1440 and 4k with a 1070 and 1080 we also
retested mordor on our gtx 980m
gtx 970m laptops that we have a home
base and also the desktop 1080 and 1070
cards on a comparable platform to the
laptops tested this is a big issue point
with desktop to notebook comparisons but
Intel's issuance of a 6700 CPU on mobile
platforms makes it much easier to
compare similar configurations with the
desktop there are imperfections in
methodology due to the limited time at
the event we only had 4 hours to work
after all but this gives a good initial
look at performance we're confident in
the results and you might not find it
elsewhere
we're also able to look at frame time so
I'll have for you momentarily here's the
1080p ultra performance chart with
average 1% lo and 0
1% low performance metrics with FPS
plotted for the new cleveo p77 5 with a
gtx 1080 we're hitting 130 8.5 FPS
average the 90 FPS 1% low and 76 0.5 FPS
0.1% low pretty similar to desktop GPUs
for the GT X 1080 overall which makes
sense since the cleveland unit has an i7
6700 at 3.4 gigahertz and so does the
test platform for the D GPUs that we
tested in this one anyway compared to
the gtx 980m that puts a 1080 laptop GPU
at more than two times the performance
output and frame rate then the 980m the
GTX 980 in laptops would be closer in
performance but our 980 unit was a loner
and we can't retest it unfortunately the
gtx 1070 and MSI's GT 62 notebook
performed at 114 FPS average 76 fps 1%
low and sixty five and a half FPS 0.1%
low the msi gtx t2 uses a lower powered
GPU than the cleveo unit at 2.6
gigahertz on the I 767 hundred HQ so the
architecture is more or less the same
and the clock rate is different the
results aren't perfectly linear in this
regard with desktop when we compare the
HQ versus the no suffix at all 6700 so
some of the doubts that can be
attributed to the CPU change with the
rest attributable to core and clock
differences in the 1070 notebook GPU
once we get these notebooks in our own
lab we can validate more and control
better for variables the gtx 1070 is
more than 2x faster than the 970 m from
these tests as we saw with a 980 m and
the GTX 1080
looking at 1440p the cleave op7 zone 5
sits at 101 dot 5 fps and 73 to 5 fps 1%
lows with a 62 fps 0.1% low the GTX 1070
notebook is about 30 FPS slower at 70
FPS and retains tightly timed at low is
compared to last gen it's clear that the
970m struggles with 1440p to a point of
the low is introducing noticeable
stutters at 92 fps and becoming less
desirable for play but the 1070 does not
have the same issue as we stated with
the Pascal and the Polaris GPU reviews
1440p appears to be entering
mainstream light and it's fully playable
on the GTX 970 1080 notebook GPUs from
the limited testing that we've done thus
far as for 4k performances about where
you'd expect it when comparing against
desktops the p7 7 5 and its GT X 1080
suggests below 60fps really not bad all
things considered it is a laptop with
the GT 60 to 1070 notebook at 45 FPS
average both units could theoretically
handle some level of 4k gaming based on
these results depending on if you're ok
with dropping settings in more intensive
games considering where we came from
with the 980 m at 26 FPS average and 970
m at 19 FPS average this is a big gain
for portable computing and at similar
power consumption frame times aren't as
tight as the desktop GPUs as you can see
in this chart of the kleva unit actually
got a few charts of it but the
performance is overall acceptable there
were no major freezes or stutters with
only a few standout points in the
overall frame time plot note that all of
these tests were conducted while plugged
in as the battery used will throttle GPU
by design for battery Boost 3.0 which is
just supposed to extend battery life by
tanking the clock rate to something like
30 or 60 FPS depending what you can
figure it to and that's all we've really
got for today we have a few of these
notebooks coming in and I'm really
looking forward to benchmarking the
thermals especially for endurance
against longer burnin periods to see how
these perform stock and overclocked and
we'll be looking at that shortly as we
did with all the desktop GPUs recently
so the same testing methodology will
apply to those once we've got them for
today though the advice is this you will
find these online they should probably
be on retailers by the time this video
goes live I don't know what the stock
will be like if they're selling out or
whatever but within normal reason it is
generally our recommendation to wait for
our review before make it a purchase
because these numbers do look good but
it was one game didn't test thermal see
I didn't have no waste ability all that
stuff so also there's a lot of vendors
out there this something like 20 laptop
vendors who are prudes for these so just
hang out a little bit we'll have reviews
online very shortly and once those hit
the web hopefully things look good and
you can make purchasing decisions based
on those numbers but for now that's it
so thank you for watching as always pay
traveling is a post roll video if you
wanna house that directly subscribe for
more content to make sure you don't miss
the next laptop video I'll see you all
next time
you
the GTX time that celebrity there need
Beauty come back
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.