Intel Haswell-E Core i7-5820K vs. Ivy Bridge Core i5-5370

This one’s been bugging me for some time, mostly because I’m an enthusiast and even more so since I started transcoding my Blu-Rays to mp4. But I’ve never pulled the trigger. The reasons are simple:

  • The GPU is not fast enough.
  • It’ll only be good for this one thing, transcoding.
  • I reasoned myself into not needing one.
  • I am a penny-pincher.

Until recently…

The last bullet on the list didn’t change, but the others did and so I finally convinced myself that it is OK to spend a few (many) 100 bucks on basically a new computer. As I’ve written in February, I replaced the aging AMD HD7870 with a brand new Nvidia GTX970 – eliminating the first item on the list.

Of course I have read many reviews and benchmarks and all show that the 6-core CPU is competing with the 700MHz higher clocked Core i7 4970K (“Devil’s Canyon”) in almost every game, or even beating it from time to time, depending on how well the game utilizes all those cores. Now, to be true, I don’t have any performance issues in games. However, recent game engines do scale pretty well.

Then there is the topic of transcoding. I don’t rip a Blu-Ray every day which would warrant this CPU, I mean, who does? But I played arouing record video games (famously dubbed “Let’s Play”; I like the format and wanted to try it). Doing this I ran into a few issues. I could either record in low quality to avoid huge files or record in a higher quality and shrink them afterwards.

The first option is not really an option. When I am watching a video I expect good quality. For some time I went with a compromise – Nvidia ShadowPlay with 15Mbit on 60 fps. That resulted in about four to five GB per file for 30 minutes of gameplay. It was tolerable for uploads but not optimal. Out of curiosity I transcoded one file with the same settings I use for my Blu-Ray rips and who’d have thought that the output file was about ¼ the size and the quality was pretty close to the input file, especially given that I wasn’t a high bitrate source. That meant I would be able store four episodes instead of just one and also upload four in the time frame of one previously. YouTube also processes those files faster because I have more influence on the parameters that are used for encoding. I can use the settings Google recommends.

The very downside of this is the time the computer needs to do the conversion. It’s between 60 to 90 minutes for about 30 minutes of video. Any of the Hobbit movies took around five to six hours if I remember correctly. During that time I couldn’t really do anything with the computer unless I wanted to hinder the transcoding process. As a result I let it run at night for a few times, but I really hate that. The economical proposition of doing this is certainly better than buying stronger hardware, but that didn’t change the fact that I hate it when my computer is running and I’m not using it.

So, I let the geek get the better of me and bought a new soul for my big tower.

  • Intel Core i7-5820K
  • ASUS X99-A
  • G.Skill 16 GB DDR4 2666MHz
  • Noctua NH-D15
  • Enermax Platimax 600 W (80-Plus Platinum)
  • Samsung 850 Evo 500 GB SSD

I could’ve probably gone with just the board, RAM, CPU and cooler but I need my computer for work as well. This way I can keep the old stuff partially assembled and do a “simple” swap in case of problems early on. I even have a spare tower to install it in. All I have to do then is swap the HDD with all the data stored on it and I’m back in the game. In addition, my two SSDs are getting a bit small and I really want a power unit with cable management (it’s also a bit stronger to accommodate for the higher power draw of the CPU). Two stones with one fly.

Let’s get to the numbers now. I stuck with the built-in benchmarks of the games and tested the highest quality in 1080p for real world performance and medium settings in 720p to take the GPU out of the equation.

Ultra 1080p Quality
Core i5 3570 (3.4 – 3.8 GHz)
Core i7 5820K (3.3 – 3.6 GHz)
Average
Minimum
Average
Minimum
Bioshock Infinite
100.0
39.0
104.6
25.1
Tomb Raider
91.5
70.0
90.0
66.0
Dirt 3 Complete
164.4
136.6
186.6
160.7
DA: Inquisition
55.9
35.4
56.7
47.8

 

Medium 720p Quality
Core i5 3570 (3.4 – 3.8 GHz)
Core i7 5820K (3.3 – 3.6 GHz)
Average
Minimum
Average
Minimum
Bioshock Infinite
184.8
47.5
224.9
22.1
Tomb Raider
308.8
218.0
470.5
352.0
Dirt 3 Complete
197.3
144.0
256.4
175.7
DA: Inquisition
104.7
77.4
177.1
119.4

In general the 6-core CPU can keep up with the slightly higher clocked Core i5 or even surpass it. What’s a little odd is the minimum FPS in Bioshock. I tested this several times just to make sure I didn’t write down the wrong number. Dragon Age Inquisition, the most recent of those games, shows a different picture. It utilizes the Frostbite Engine that’s also responsible for pushing pixels in Battlefield and clearly shows, especially when the GPU is not the limiting factor any more, what more cores can do for you.

Now the transcoding benchmarks. I have used the x264 HD Benchmark (32 Bit, that’s what I had installed) and Handbrake (64 Bit). For the latter I took a 2.36 GB big file (13 minutes) and encoded it with the settings I use for YouTube.

Core i5 3570 (3.4 – 3.8 GHz)
Core i7 5820K (3.3 – 3.6 GHz)
x264 HD Pass 1
40.6 fps
83.1
x264 HD Pass 2
11.4 fps
21.7
Handbrake
28m 46s
15m 17s
C++ Compile
8m 35s
6m 55s

Twice as fast. The i7 5820K destroys the Core i5 when it comes to video transcoding. After all, there are two more physical cores and six more threads due to hyper-threading, something the i5 doesn’t even have. I’d say it is a comparison between a quad-core and an octa-core (I postulate that three virtual cores are about one physical). I also threw in a complete compilation of one of the applications I work on in the company I’m employed at. It’s a bit faster but not much. I suspect the HDD is not able to deal with the additional load. Nevertheless it is about 90 seconds quicker now.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s