Are All MHz Created Equal? - Intel 5960X vs Q6600 Comparison

Linus Tech Tips ·Linus Tech Tips ·2015-05-07 · 1,373 words · ~6 min read
Floatplane YouTube

Transcript

JSON SRT VTT 99
0:00 There's an age-old question that gets asked all the time. Are all mehz the
0:04 same? Let's talk about that.
0:15 The Cooler Master Novaouch TKL utilizes genuine Toper hybrid capacitive switches
0:20 and is now available at a more affordable price. Click now to learn
0:24 more. Good tech tips morning. I love
0:28 GMM, but what we're here to talk today about is the myth of megahertz, not the
0:33 awesome antics of Rhett and link. So, let's set the stage. There are many
0:37 things that will determine a processor's effective speed when handling tasks. But
0:41 two of the most common ones are clock speed, which is a given, and IPC, also
0:47 known as instructions per clock, which you'll hear a lot when it's thrown
0:51 around uh as the banner of IPC improvements with every new processor
0:56 launch. Right about now is when I would get you guys to click on the box to go
0:59 check out the fastest possible episode explaining IPC, but we don't have one as
1:05 of yet. So, what I would suggest is you check out the kind of somewhat rough
1:08 Wikipedia article on the topic and other various sources. We'll try to get a
1:12 video about that out fairly soon. IPC isn't necessarily just as simple as it
1:17 sounds at first, but Intel claims generation after generation to have a
1:21 usually around 10% general improvement. Does this improvement actually amount to
1:26 anything? Wouldn't an old processor running at 2.4 GHz run just as well as a
1:30 new processor running at 2.4 gig gigahertz? And even if the improvements
1:34 do amount to something, will this show up in non-realworld synthetic tests
1:39 only, or will it actually be noticed? Let's find out. The tools of the trade
1:44 for today will be two different test machines. One will be Lionus' Scrapyard
1:47 Wars machine. You haven't seen Scrapyard Wars, first of all, I'm genuinely
1:51 surprised. And second of all, click here or something and then come back later.
1:55 We will be using the Q6600 that which is
1:58 running in his system at 2.4 GHz rocking four standard cores and no
2:02 hyperthreading alongside 8 GB of RAM and a freaking GTX 980 from uh Gigabyte
2:08 Windforce just to push the GPU bottleneck out of the way and see how
2:11 effective the processor actually is at handling things. Our second system will
2:15 be the standard GPU test bench machine with its smuggly running 5960X, but
2:20 today it won't be running normally. I've nerfed the 5960X down to 2.4 GHz and
2:25 four cores with no hyperthreading to match the clock speed and core count of
2:29 the Q6600. If you're interested in how core count can affect things, check out
2:33 my recent video here on cores for gaming. Then I slapped 8 GB of RAM onto
2:38 the bench and the same Windforce GTX 980 to even things up. In order to pit these
2:43 two systems against each other, I picked a suite of seven different benchmarks,
2:46 three different synthetics, and four games. I'll start up with a forum
2:50 favorite, and that would be Cinebench R15. Check out the awesome Cinebench R15
2:55 thread and other RS actually on the forum. As you can see, there's been a
2:59 whopping 55% improvement. That's pretty
3:02 huge, even given the 7-year gap between these two processors. And with our new
3:07 fancy new chip majorly nerfed just for
3:10 this experiment. 3D Mark and 7zip tests were fairly similar with the 5960X
3:15 actually slapping around the old weary Q6600. But none of this was all that
3:19 surprising. But it's time to move on to stuff that you're probably more interested. Real world gaming
3:24 performance. First up, we have Tomb Raider, which showed essentially no
3:28 difference between the two. I really need to stop using Tomb Raider for
3:32 anything other than GPU tests because it really doesn't scale with anything other
3:36 than GPUs at all. Well, that story was
3:39 boring. So, moving on. Next up, we have Far Cry 4, which as we know from the
3:43 course for gaming video, isn't all that CPUbound, unless you count not running
3:48 on dual cores. But regardless, we do see a bit of an improvement here, and that's
3:52 nothing to scoff at. But this is where it starts getting interesting. dying
3:56 light. Fairly substantial difference in average FPS between the Q6600 dipping
4:00 all the way down to 15 FPS at times for minimums and the 5960X holding a cool 25
4:05 FPS for minimums. Also note that whenever there was a slightly more
4:09 intense zombie action going on, it would chug a bit and even crashed a few times
4:13 on the Q6600. Not a great experience.
4:16 Last of all, and oh did I save the best for last, is City's Skylines. We just
4:21 recently introduced this game into our testing suite, and damn does it love
4:24 processors, and it really shows here with the Q6600. There were stutters all
4:30 freaking over the place. It couldn't run 4K, and it was barely able to manage
4:34 itself at 1080p, and it was just a horrible experience. Meanwhile, the
4:39 5960X was just chilling. No big deal.
4:42 So, in conclusion, what does all this mean? Well, I hope it helped to show how
4:46 things can improve more drastically over time than most the easyto read specs
4:50 actually show. If you're already happy with your late gen processor, you don't
4:54 need to run out and replace it this second. That's doesn't matter. There are
4:58 always other things that you can do as well. Overclocking can help a ton and
5:02 you can just not run all your games at max like we just did. That can help too.
5:06 But if you are a speed freak and you've been struggling to figure out why your
5:10 older overclocked processor isn't performing quite as well as your buddy's
5:14 stock processor running at the same frequencies, hopefully this will help.
5:19 Speaking of better than dessert, Crunchyroll. If you're a fan of anime
5:22 and haven't signed up for Crunchyroll Premium yet, you should certainly give
5:25 them a shot. Crunchyroll is a site created by anime fans for the other
5:29 anime fans. They offer the most current episodes of new shows straight from
5:33 Japan, like Kuroko's Basketball 3 and Jojo's Bizarre Adventure, as well as a
5:38 large collection of the most popular anime shows like One Piece and Nudo, all
5:43 professionally subtitled. If you head over to crunchyroll.com/Linus,
5:48 you can sign up for a 30-day free trial of Crunchyroll Premium, which is
5:52 completely adfree. And if you enjoy many benefits of premium, like 1080p
5:56 streaming, getting new episodes of shows straight from Japan within an hour of
6:00 their premiere, and being able to stream anywhere, anytime from a variety of
6:04 devices, even your phone, tablet, or heck, Wii U, you can continue your
6:10 premium membership to Crunchyroll for only $6.95 per month. So, head over to
6:14 crunchyroll.com/Linus and check them out.
6:18 Personally, I'm quite a fan of my 5 GHz 2600K, but it may be time to upgrade
6:24 soon. Let me know what you think down in the comments down below, or better yet,
6:28 over on the forums. While you're here, like, dislike, favorite, subscribe,
6:32 share, do all that kind of stuff. Check the link in the video description down
6:36 below to get a cool Linus Media Group shirt. There's a bunch of really cool
6:39 ones, and I actually have been running into people wearing them at conventions lately, which is pretty sick. Over on
6:44 the forum, you can check out the support Linus Media Group link when you where you can do cool stuff like changing your
6:48 Amazon affiliate code to ours. That gives us a small kickback whenever you buy things and actually adds up to being
6:54 really really helpful. Also, if you want to become a contributor on the forum,
6:57 you can get a really cool little badge under your profile picture whenever you
7:00 post things. Thanks for watching, guys, and I'll see you next time.