PUBG CPU Benchmark, Very Low/Medium/Ultra Results [16 CPUs Tested]
PUBG CPU Benchmark, Very Low/Medium/Ultra Results [16 CPUs Tested]
2017-12-03
welcome back to harbor unboxed today
we're looking at cpu performance in
player unknowns battlegrounds now about
6 months ago we did put together a mini
tests that compare GPU performance using
the games of various visual quality
presets and found that the best value
combo for those targeting 60 FPS at Sony
P was the rosin 514 hundred and GTX 1060
since in the game scene Calais updates a
number of them have addressed the games
proteome ization issues or at least
they've claimed to that said in its
current state the game still requires a
lot of work things have been improved
but I'd say overall only a little bit
since we first tested it all those
months ago roughly four months ago
though it did receive a patch which
improved CPU utilization claiming a damn
can now utilize six or more cause over
the past few months I've been stuck with
the game a bit but I haven't done any
in-depth testing where I compared a
large number of CPUs it's mostly been
head-to-head type stuff so I thought why
not take all the eighth generation Intel
Core Series CPUs all the rise and CPUs
and a few seventh gen core CPUs and test
them in pub G under the same conditions
this is exactly what I've ended up doing
and now we have results for sixteen
different CPUs at 1080p using the very
low medium and ultra quality presets
with the GeForce GTX 1080i which has the
380 8.43 driver installed all unlocked
Intel CPUs along with all the rise in
CPUs have been tested using ddr4 3200
cell 14 memory meanwhile the locked
Intel CPUs were a test with ddr4 2400
cell 14 memory so for example the core
i3
8350 K was tested with 3200 memory but
the core i3 8100 was tested using 2400
memory towards the end of the video I've
also noted the CP utilization of all 16
CPUs for those of you interested and
there are some interesting results to be
seen there for testing I'll walk through
one of the busy towns for 30 seconds as
this is more than enough time to gather
the data we need who actually reduced
the past time from the normal 60 seconds
just 30 seconds to try and minimize the
frequency at which I was killed before
completing the pass as this is a high
loop area so higher
risk but high reward and yet high risk
for those benchmarking as you would
expect the test starts at exactly the
same point and ends at the same point
every single time and I take an average
of three runs so let's check out their
results
first up we have the very low quality
results here the visual quality settings
are set to their lowest value so this
should remove the GeForce GTX 1080i as
the performance limiting component that
said we are clearly seeing a GPU
bottleneck here with the majority of the
seventh gen and 8th gen core processors
the gtx 980ti looks to be good for only
about a hundred and twenty FPS on
average with dips to about a hundred fps
previously when testing with the ultra
quality preset I found the seventy-seven
ok and r5 1600 for example delivered the
same performance here we see though that
the seventy-seven ok is 20% faster than
the 1600 X as the Rison cpus appear to
be struggling in comparison of course
with well over 60 fps at all times the
Rison cpus did still provide very
planned performance but in a game that
claims support for high coil count CPUs
the results are disappointing on that
note I will discuss CPU utilization
shortly quite shockingly the horizon 7
1800 X was just 14 percent faster than
the Rison 3 1200 for the average frame
rate and just 6 percent faster than the
1300 X so this suggests that the game
really isn't utilising higher core count
CPUs very well at all and instead
prefers core frequency / core count
based on these results that appears as
though a quad core really is enough and
it doesn't even necessarily require HT
or SMT support that said though for
optimal performance a dual core with HT
n able isn't enough and we see this with
the Pentium G 4560 which really was
considerably slower than any other CPU
tested that said it was still able to
provide playable performance and would
be a great pairing for a low-end $100
u.s. graphics card increasing the visual
quality settings with the medium quality
preset reduces the GTX 1082 s
performance by around 10 percent with
the faster CPUs tested there is a little
bit of reshuffling with the horizon cpus
and now those with more cores are seen
to be doing a little bit better at least
when compared to what we saw previously
the 18
Nexxus now 25% faster than the twelve
hundred and sixteen percent faster than
the thirteen hundred x so the medium
quality settings do appear to place a
bit more load on the CPU though I have
to say this wasn't apparent when
monitoring CPU utilization as the
overall figures were much the same again
we see that the Intel CPUs are able to
find the limits of the GeForce GTX 1080i
so it's likely the eight and twelve
threader models could go faster again
finally we have the ultra quality preset
results and here we see very little
change from the medium quality results
for the most part just a few frames are
dropped though it's the Intel quad cores
that are the biggest losers here the one
percent low result for the core i3 80
183 50k dropped by around fifteen
percent whereas the Rison five 1500 X
and rise in 313 hundreds for example
were just eight percent slower has to be
said that so far the results seem to be
a bit all over the place and that's
something we do often see with poorly
optimized games this graph gives us a
better look at what's going on so it
doesn't really clear anything up looking
at the core i7 a7 okay we see a 10% drop
from the very low to medium quality
presets and then a 3% drop from medium
to ultra the core i5 7600 K on the other
hand drops 10% from the very low to
medium preset and then a further 12%
from medium to ultra so that's
interesting the ultra quality preset
certainly hurts the quad-core more but
then we have the horizon 3 1200 results
which are more in line with what we saw
from the 8700 cave so that's confusing
to say the least
then the 1600 X shows fairly consistent
scaling across the three quality presets
then we have the pentium g 45 60 and
that's different again showing similar
results with the very low and medium
presets and then dropping quite a bit
when using the ultra preset so overall
we see at the core i5 7600 K and Rison 5
1600 the only CPUs to show consistent
scaling okay so last that we have the
CPU utilization results and they are
very interesting firstly what you're
looking at here is the average CPU
utilization recorded from our 32nd past
so it's not the peak but rather the
average the G 45 60 for example did at
times hit 100% but it also dropped down
to around 80% and for the entire test we
did see an average utilization figure
of 91% what's interesting to note here
though is how heavily underutilized the
risin 5 and risin 7-series CPUs are the
horizon 5 1600 X for example has 4 cores
and 8 threads and it's clocked it up to
3.7 gigahertz depending on the workload
yet the core i7 87 are okay which also
packs six cords with 12 threads and a
minimum operating frequency of 3.7
gigahertz actually saw a higher
utilization quite a bit higher in fact
AMD zone 6 core 12,000 five sixteen
hundreds or an average utilization
figure of just twenty eight percent
which is considerably lower than the
forty five percent figure seen when
testing with the 87 or okay you would
expect a lower clock CPU the same core
in thread count to see higher
utilization not drastically lower so
there is clearly a serious optimization
issue here for rise and CPUs six months
on and player unknowns battlegrounds
still requires serious optimization work
shaadi rise and support aside even the
core i7 eighty seven okay and gtx 1080i
combo were very underwhelming pushing
just a hundred and twenty three FPS on
average tally pair using the minimum
quality preset well for a lack of a
better word and that's pretty pathetic
helping to put that result into context
the same combo pushes over two hundred
and twenty fps and battlefield one using
the medium quality preset we see 200 FPS
in warhammer to 240 FPS in f1 2017 260
fps and Rainbow six siege 220 fps and
Call of Duty World War two and while the
list just goes on again with all those
games we're nowhere near the minimum
quality settings either all quoted frame
rates are based on the medium quality
preset in anticipation that some allow
you that pub G is an open-world shooter
and therefore hammers the CPU well as we
saw when testing with the very low
quality settings this simply isn't the
case an Intel quad core or grader will
hit a GPU bottleneck at just a hundred
and twenty FPS while whole horizon CPUs
were heavily underutilized and didn't
have a chance to get anywhere near that
figure one thing seems clear if you're a
massive pub G fan and you're building a
new computer solely play this title
something like the core i3 81
or a 350 K for example will offer you
the most bang for your buck
I'd never normally recommend the 80 350
K but pub G makes that somewhat have a
valid choice although I am yet to test
any older CPUs based on the results that
we have seen here I would quite
confidently say that anything back to a
core i5 2500 K will play the game just
fine providing an overclock is applied
and it's pretty shocking to find though
on the GPU front at least with a high
end GPU that there is very little
difference between the very low and
ultra quality settings in terms of FPS
performance visually though there is a
massive difference so again that's quite
shocking
anyway that's gonna do it for this one
let me know what you think below as
always I'm keen to hear your thoughts
I'm your host Steve see you next time
you
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.