AMD A8-3850 vs Phenom II X6 1100T With Radeon 6990 Linus Tech Tips

Linus Tech Tips ·Linus Tech Tips ·2012-05-07 · 1,083 words · ~5 min read
Floatplane YouTube

Transcript

JSON SRT VTT 81
0:07 All right, guys. I got a good one for you. I'm going to take the Phantom 2X6
0:11 1100T and I'm going to take two Radeon
0:15 HD 6970s. Okay. Yes. Yes. This is a
0:18 6950, but just to prove that I have
0:21 successfully unlocked the card. You can see here both GPUs are identical under
0:26 GPUZ. So, I do have them performing at 100% 6970 speeds. And I'm going to be
0:33 comparing the 1100T on the 890 FX
0:36 platform. See, I've got a Crosshair 4 formula there. And I'm going to compare
0:40 it against the A75 platform
0:44 and AMD's new A8 APU. So, I'm going to
0:48 find out what kind of a performance difference there is between their new
0:52 APU platform, which does have slightly higher performance per clock than at
0:57 least the Athlon series. So, it should be pretty close to a Phenom 2, okay, in
1:02 terms of clock forc clock performance, but this is a quad core. And on the
1:06 higherend platforms, you can get six cores now and eight cores coming soon.
1:09 So, let's see how this goes. I'm going to be testing a variety of current
1:12 titles including Crisis 2, DirectX11,
1:15 Witcher 2, Dirt 3, Fear 3, Battlefield Bad Company 2, and Civ 5. Don't mind the
1:21 other things on here, that's going to be for something else. Small change to the
1:26 methodology, guys. I'm going to be using my 6990 because Crisis 2, apparently, in
1:31 spite of using the 11.6 six drivers, the latest uh profiles, and the latest patch
1:36 of Crisis 2, including DirectX11 and high-res textures.
1:41 Still doesn't work in Crossfire. Go figure. So, these are all the runs of
1:45 Crisis 2 I did trying to go. Okay, single card, multicard, single card. Oh,
1:51 you guys can't see the results that I'm bringing up. So, you can see that we're
1:55 getting some pretty good consistency there. In spite of the fact that I was
1:58 switching between a single 6T970 and dual 6T970. So, I throw the 6T990 on
2:04 there. All of a sudden, I get 49.5 FPS.
2:07 So, we're going to stick with that for our testing. So, I'll be testing that against the 1100T or with the 1100T
2:12 against the A75 platform. So, I've got
2:16 myself a benchmarking snack bowl. Unrelated to that though, I have
2:20 finished running my tests for the 1100T which was here. So that's running on the
2:26 Crosshair 4 formula with the 6T990. And then on my other platform, I've got the
2:31 AMD Lano A83850 running with exactly the same test bench
2:36 with the 6990. So, I ran a variety of modern games, and what I found is that
2:41 with an identical configuration. Remember, with the APU, we're only
2:46 taking advantage of the CPU component of it because the onboard graphics turns
2:51 off if you install anything higherend
2:54 than about a shoot, what is this? Turks, I think it's 6670.
3:00 So, if you install anything higherend than something that looks about like
3:03 this, it is going to turn itself off because it's not going to provide any
3:07 additional benefit in dual graphics mode. Okay. So, we're looking at its
3:12 performance strictly as a gaming CPU against the 1100T, which is AMD's cream
3:18 of the cream. Now, I think you'll find that in terms of price toerformance
3:21 ratio, it's pretty darn impressive. So, the games I tested were Crisis 2. This
3:26 is DirectX11 with the high resolution texture pack. So, everything is maxed
3:30 out in this game. In fact, if I bring up
3:33 my little email that I made for myself outlining all of my benchmarking
3:37 settings. Uh, yeah, there we go. Okay, I don't have Crisis 2 in there, but whatever. It's It's all maximum uh with
3:43 the DirectX11 in the high-res pack. Actually, no, I think I'm at Okay, give
3:47 me a sec. Okay, sorry guys. For Crisis 2, I'm using all extreme presets and
3:52 then um I'm running at obviously 1920x
3:56 1080 and DirectX11 is enabled and the uh
4:00 uber high-res textures are enabled. So for Crisis 2, the 1100T performed about
4:06 10% faster than the A8 APU with exactly
4:09 the same configuration. The Witcher 2 was much closer. It was only about a 5%
4:13 difference, which is pretty much, I'd say, within the margin of error because
4:17 I am doing game runthroughs. I am not doing canned benchmarks on Crisis 2,
4:22 Witcher 2, Fear 3, and Battlefield Bad Company 2. So, there is some variance,
4:26 although I do my best to run through exactly the same way every time. Now,
4:30 Dirt 3 was one where the 1100T just blew
4:34 away the A8PU. It performed about 30%
4:37 better. So, 25 30% better. And Fear 3
4:41 was one where they were very close. So, we start to see a pattern here. Uh, come
4:44 in and have a closer look at the graph. So, we start to see a pattern here that says 76.2, 2 by the way, sorry guys,
4:49 where any game that is CPUbound or very multi-core aware is going to perform
4:54 better on the 1100T platform, whereas any game where we're already quite
4:58 graphically bound, such as The Witcher 2 or Fear 3, we're not going to see much
5:02 of a difference in performance. Battlefield Bad Company 2, um, perhaps
5:06 due to the increased CPU, uh, load in that game, it does a lot of physics
5:11 effects on the CPU, performed about 20 22% better on the 1100T. And finally,
5:17 Civ 5 performed about 10% better. And Civ 5, I was using a late game view uh
5:23 of a game that I actually played. I have about 40 cities. So, what I do is I
5:28 center it over one of my cities and then pull out all the way and see what the
5:31 frame rate is in the corner. So, thank you guys for checking out my little
5:35 gaming performance showdown between AMD's highestend CPU and then their new
5:42 just phenomenal, although I can't really call it phenomenal because it's not even
5:46 a Phenom. Um, wa that's mind-blowing,
5:49 right? And their new APU. So, stay tuned for more APU content around the 3850.
5:54 And don't forget to subscribe to Linus Tech Tips for more unboxings, reviews,
5:57 and other computer