AMD RYZEN 5 AND 3 PERFORMANCE PREVIEW!

Linus Tech Tips ·Linus Tech Tips ·2017-05-06 · 1,261 words · ~6 min read
Floatplane YouTube

Transcript

JSON SRT VTT 119
0:00 guys guys you won't believe where I've
0:03 been or when I've returned from the far
0:08 off time of Q3 2017 after the launches of AMD's ryzen 5
0:14 and ryzen 3 and you don't want to know
0:18 what happens between now and then but anyway that's not important right now
0:23 what is important is I've brought back a
0:26 little souvenir Welcome to our ear early
0:30 preview of ryzen 3 and ryzen
0:42 5 the Phoenix oour is a full-sized
0:45 minimalistically designed keyboard complete with Cherry MX switches and a
0:49 new white backlit variant check it out at the link
0:53 below okay okay so I'm no do brown and I
0:58 drive a Lambo not a DeLorean so we
1:02 should just get this out of the way right now we don't actually have a ryzen
1:08 5 or a ryzen 3
1:11 chip or do we no no we don't this right
1:17 here is our ryzen 7 1800x and what we're going to do is use
1:23 our Motherboard BIOS to match the core
1:26 counts and frequencies that AMD has
1:30 announced and see how they might stack
1:34 up you may have your doubts about this
1:37 method but we've demonstrated in the past that it works shockingly well in
1:43 this video here so which chips will we
1:47 be testing then well aside from our ryzen 7 1800x and
1:53 1700x as real world baselines and a core
1:57 i7 7700 K to rep present the blue team
2:01 we're taking the ryzen 5600x and 1400x and the ryzen 3 1200X
2:09 Forest spin we'll be doing it all with
2:12 our replacement ASUS crossair 6-based
2:16 ryzen test bench and as for our GPU
2:20 we'll be testing with our new FPS per dollar high performance King the GTX
2:25 1080 TI now considering that the highest
2:29 clocked r 5 CPU is still a 3.6 GHz base
2:34 4 GHz boost chip the same as the 1800x
2:38 we're pretty certain that we're not going to see any performance difference
2:42 in single-threaded tasks with the 1600 X
2:46 making it slower in multi-threaded workloads but potentially a wicked
2:51 gaming chip there are some things to
2:54 keep in mind with these tests as AMD's
2:58 Precision boost is disable AED the moment a custom frequency is set we've
3:03 had to lock the frequencies of our slower clocked virtual CPUs since
3:09 Precision boost and xfr only kicked in
3:12 in lightly threaded workloads in our testing we chose to split the difference
3:17 between their base frequencies and maximum xfr clocks that is 3.75 GHz for
3:23 the 1400x and 3.6 GHz for the
3:27 1200X this clearly isn't 100 100%
3:30 accurate particularly since that puts our lower end chips running at a higher
3:34 base frequency while at full loads it
3:37 should still give us a ballpark representation for our overall
3:42 performance analysis across all of our tests finally we're uncertain at this
3:48 time whether disabling cores on ryzen
3:52 also disables extra cash so do keep that
3:56 in mind with all that boring stuff out of the way now though let's kick it off
4:01 with our gaming benchmarks today at 1080p since these are mid-range CPUs
4:07 that we're simulating we see our virtual 1600x performing better in DSX mankind
4:14 divided than the lower clocked but pricier
4:18 1700x rise of the Tomb Raider shows us more of the same while for Honor shows
4:23 our, 1400x also performing well with its
4:27 higher base frequency sadly an artifact of our methodology
4:32 since ryzen doesn't boost for that long at full load GTA 5 clearly will take all
4:39 the course and Cycles it can get while Ghost Recon wildlands on the other hand
4:44 shows only the i7 7700 K's
4:47 single-threaded performance pulling away with a measurable lead Doom gives the
4:53 7700k a much more significant lead with the ryzen series making a pretty linear
4:59 slope at after that and finally like GTA 5 Crisis 3 shows again a thirst for
5:04 everything with our 1200 X falling far behind thanks to its mere four cores and
5:10 four threads moving on to synthetics 3D
5:14 Mark shows our ryzen platform generally outperforming the
5:18 7700k until we get all the way to our quad course and the same is true of szip
5:24 where even our 1600x outperforms its
5:27 more expensive Intel rival PC Mark is
5:31 another highlight for AMD the 1600 X's performance is close to the 1800x and
5:37 another instance where our methodology benefited our 1400x in a way that won't
5:42 quite hold true in the real world the
5:45 1600x again pulls ahead in the single-threaded Y cruncher test but our
5:50 multi-threaded test brings us closer to
5:54 a kind of linear slope cinebench 2 shows
5:57 us fairly predictable results and except for the i7 which only beats the much
6:02 cheaper quadcore ryzen chips rounding
6:06 our test out ASUS realbench shows us its
6:09 preference for clock speed for image editing thread count on encoding and a
6:14 combination of the two in heavy multitasking with our benches stacking
6:18 up accordingly but to put all of that
6:22 performance stuff in the appropriate perspective we need to look at pricing
6:28 at24 9 the 1600x seems to hit the sweet
6:33 spot with its 1800x like performance in
6:38 everything but our most multi-threaded
6:41 benchmarks the 1400x at 169 seems like a
6:46 really solid value too considering its
6:49 thread parody with our i7 and decent
6:53 scoring which lands it at the top of the pack for performance per dollar the
6:58 1200X rumored at 149 is uh well AMD's
7:04 marketing is all about thread count and with only four ryzen just isn't pulling
7:09 quite as much weight however at that price it's actually our second best
7:15 performer per dollar placing it in an excellent position for a budget CPU and
7:22 if you've watched our ryzen OC guide you're probably thinking that this one
7:27 is actually much closer to the ath XP 2500 plus than the
7:32 1700x and you wouldn't necessarily be wrong though something our 8 core
7:38 processor cannot simulate is how hot these chips will run or how well they'll
7:44 overclock so that all remains to be
7:47 seen so with all of that said then we have to stress again that this isn't an
7:52 exact measurement in particular our 1400x and 1200X scores are likely
7:58 suffering from the loss of PR ision boost but taken with a grain of salt
8:03 these results should paint a pretty clear and pretty exciting picture one
8:08 that we can't wait to verify against the real
8:12 deal Zotac magnus1 series is one of the
8:17 smallest VR ready gaming PCs available
8:21 measuring at just 8.27 by7 by 2.45 in
8:26 it's quiet while operating and it's light we at less than 8 lb making it
8:31 genuinely portable you can even get it with an NVIDIA GTX
8:35 1070 to go along with its skylet Core i5
8:39 6400t its dual HDMI 2.0 ports and its
8:43 compatibility for high-speed NVMe ssds
8:47 you can learn more at the link in the video description and get your
8:53 own so thanks for watching guys if this video suck dislike but if it was good
8:57 then like it get subscribed for more videos like it and check out the link to
9:01 where to buy the stuff we featured in the video description also down there is
9:06 a link to our merch store and our community Forum which you should totally
9:10 join