Wow.... 53 watts more than the 8 core Ryzen... can we start with the Intel hotter than the sun jokes?
per core its a little faster (like 5%) than the 1800X for a bucket load more watts.
-
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
AMD THREADRIPPER VS INTEL SKYLAKE X
- Thread starter faceman123
- Start date
Wow.... 53 watts more than the 8 core Ryzen... can we start with the Intel hotter than the sun jokes?
per core its a little faster (like 5%) than the 1800X for a bucket load more watts.
~43% increase in performance for a ~62% increase in power not really unsurprising as its on the same node and comes from increasing frequency and hence voltage - pretty much the expected ratio for increased power with voltage.
Works out around 20 watt per core for both and Intel has a slight performance advantage per core.
the 10% (nobody said 30%) ipc lead is still there, guess you failed to notice how at identical clock speeds (4ghz 1800x vs 4ghz 7900x) the Intel had nearly identical power draw, that means this process is more efficient than amds clearly.
anyway, this shows what we were expecting, decent and easily overclockability, very good single core performance and very good multicore performance, the only outliners are obviously as bit tech said, a bios or software issue.
hexus has a review too, apparently they got a bios update today (half way through their tests) that showed a large uplift in some of the odd results.
guessing the board manufacture aren't ready yet as according to both hexus and bit tech neither of them actually got review sample boards.
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,335
- Location
- Planet Earth
Its the first time in yonks I have seen a new uarch Intel CPU actually have significantly worse performance/watt over its predecessor.
Yeah,but the Intel CPU is on a better process and the R7 1800X has significantly worse performance/watt than say a Ryzen 7 1700.
Handbrake looks to be 26% faster than an 1800X at stock and 42% faster when at 4.6 GHz. Remember that to calculate how much "faster" one CPU is compared to another, you need to take the reciprocal of the "time taken" and then divide one by the other. In this case, (1/60)/(1/76) = ~1.27.
Essentially Intel pulls ahead because of its higher maximum overclock.
29% ^^^^ 87 FPS + 29% = 113 FPS, this for an 8 core Ryzen vs 10 core 7900X
Cat was comparing to the 6950. Which is where my numbers are worked from.
I'm comparing it to the 1800X ^^^^
Its also a high end 10 core vs AMD's mainstream 8 core, and £700 more expensive, wait for threadripper, in think if the Intel fanboys were to be honest they would say there were expecting it to be a lot faster than that, it absolutely depends on its overclocking ability because out of the box thread per thread its little to no faster than the 1800X.
Its also a high end 10 core vs AMD's mainstream 8 core, and £700 more expensive, wait for threadripper, in think if the Intel fanboys were to be honest they would say there were expecting it to be a lot faster than that, it absolutely depends on its overclocking ability because out of the box thread per thread its little to no faster than the 1800X.
I was talking in the context of Cat's post that I replied to - so your post wasn't really very relevant in reply to mine.
Often new uarchs have either been on a node shrink, half-node or the first run off a revised process - this is just another respin of the already revised process node so given how the increase in voltage/frequency works with silicon in terms of power efficency its pretty much an expectable result. There is pretty much nowhere for Intel to go until 10nm.
Certainly on price for performance, i think, Threaripper will rip Skylake-X a new one, AMD's $850 32 threader will destroy this 7900X in terms of productivity performance. its going to be a humiliation for Intel.
source on this threadripper prices?
Caporegime
- Joined
- 9 Nov 2009
- Posts
- 25,335
- Location
- Planet Earth
Whats interesting is if you take per thread performance into consideration Ryzen at least for Handbrake and Cinebench(and even most of the non-gaming tests on BT) is doing surprisingly well considering its a lower clocked chip and unlike previous AMD CPUs is doing so at reasonable power consumption,and its only running half the memory channels too.
Its still short the 7900X though I think? bit simplistic but given the performance per core if you added 2 cores to the 1800X you are still almost 4% short of the 7900X? and based on a horrid nasty use of a coefficient representing the efficiency of the process at higher voltage/frequency or core count the 1800X would be at almost 230watt or worse at the same performance as the 7900X.
Why would anyone expect it to be any faster? We already know the IPC, it's a variant of Skylake/Kaby Lake. We already knew the rough overclocks we could expect too.
Associate
- Joined
- 12 Jul 2015
- Posts
- 1,694
Delete please.
Last edited:
Well, you know it doesn't work like that, the 4 core 7700K pulls around the same power as the 1800K, so Skylake-X having 150% more cores should pull 150% more power right?
No, and it doesn't. nor will a 16 core Ryzen use 2X as much power as the 1800X
Besides all of that the total system power here for the 7900K is 214 watts, 161 Watts for the 1800X, so on the Intel side the CPU alone is using 53 watts more, the 1800X we know uses about 90 - 95 Watt in Handbreak, that's 55% more power for 20 more threads on the 7900K.
Thats not good...
or rather, a 4ghz 8 core ryzen draws the same power as a 4ghz 10 core intel..
True it's on the same node (although supposedly the node is meant to be better, hence they call it 14nm+ now), but Skylake is meant to be a new arch vs Broadwell. So they've not even managed to keep linear performance per watt by upgrading the arch? (i.e. the arch can't compensate for the increased voltage, and you'd expect it to help a tad)
EDIT: WAIT, hold the phone, Skylake-X at 4.6 GHz is SLOWER than Broadwell-E at 4.4 GHz in gaming? What's happened there?
Whole point of my post which apparently went over your head was pointing out that it isn't a linear progression. Comparing against other configurations becomes complex as you have to take into account the potential implications of 1 or multiple dies, etc. you also have the problem that if you have two dies with say 8 cores but 4 disabled versus a die that is designed with 4 cores from the ground up and no dark silicon that can completely change the power properties and thermal dissipation, etc. generally though a generic increase in area of a die or voltage/frequency the worse the performance/watt becomes.
Looking at the 7900X reviews it seems that frequency for frequency (not sure on voltage) they are pretty much the same performance per watt i.e.
6950X @ 4.4GHz 368watt system power
7900X @ 4.6GHz 378watt system power