4770k vs 8350 for gaming and video rendering

hunterj

Honorable
Jan 4, 2014
151
0
10,710
Is the 4770k worth the extra money over the 8350? I understand that games may start using the extra cores of the 8350 but is the 4770k still better in the long run? Either way it is going to have a asus gtx 770 and a rog series mobo
 
Solution


If its straight gaming then i5 would be plenty. If theres any rendering or intensive programs at all the price difference is more than worth the i7. As for what else I would do or spend money on, since you already have gpu, maybe an ssd if you don't. Otherwise, it'd be cash for a video game, system wise hero board is top end. The formula is nice but doesn't add anything close to being worth $100 more. The extreme isn't even...
depends what your gonna be using for the renderer. if your thinking 3ds max then your gpu will be doing the work if you go with nvidia, if you went for an amd gpu you can still do gpu render but its not as well supported and has to be downloaded separately as an add in.
rendering with the cpu is comparatively slow compared to gpu rendering but the 4770k would do the job quicker at lower clock speeds (you would have to push the 8350 to 5+ghz to match 3.9ghz on the intel).
 
Its really how much do you want a boost? The 8350 is almost as good as the i7. But the mobo for asus can be had for a bit cheaper as well. Is the intel worth the money, yes. But with the added (average $100) to go intel route, you can always use that elsewhere. Which is what I did. I originally spent that extra $100 on gpu. But you can always use extra $100. Whether it be better GPU, ssd, extra hhd, spending on custom loop, nas server, theres always money to spend on stuff. Is the i7 better than the 8350, short yes it is, is it $50-100 better? That's open to personal preference. I don't see it as $100 better, but it will save you electricity which in 3-5 years could make up the difference.
 

i would say most of this is great info. but i have to disagree with the is it 100 better.
the difference in render time off the cpu can be substantial the higher you go with poly count and render quality.
i run 3dsmax to do things like this

and at 1080p it takes about 2 hours to render off the cpu if i was using an amd cpu i would be looking at maybe 2.5-2.7 hours for the same. over a year that can result in a substantial time saving and as we all know time is money.

however i do agree that if he switches to a gpu rendered he could save money with the 8350 and put it towards a better gpu which would give even more of a return.
 
well I've jumped the AMD boat and built my first intel . I went from the AM3+ to a 15 4670 and I cant say I'm disappointed now I wonder if I would have spent the extra on that i7 or not just knowing how much I been liking this i5. but to each his own.. but yes I feel this i5 is doing better that amd and it would be hard to go back to them unless they pull a rabbit out of there hat with a new , better , chip and a refreshed chip set.. that 990fx it what now? 4 years old amd is past due for a more up to date board with an improved chipset like the forgot about and dead 1090fx. but that's just me .
 


Theres no reason not to go with gpu rendering. CPU still gets 90-100% utilized, but not dependent with gpu rendering. Most people would have a gpu reguarless (gaming or something) where its not normally an added expense.

I say this, as this was the foundation on my $100 better. Yes I realize that 100% cpu rendering intel does edge out (not sure if its by as much as you stated, but I haven't done any testing to know otherwise (straight cpu rendering)). But this is why I state that either way its gong to be personal decision.

The differences between the two are minimal to decent between the two depending what you do, but even a $100 gpu if used for acceleration would greatly beat out straight i7, if used with 8350. Adding $100 to your gpu budget might not net a huge increase compared to the original gpu you were to go with, one cavat this way. But it could get you a larger ssd for boot, a second ssd for data, another hhd to add in a raid, a nas server, a lot of possibilities. And that list can keep going, and that extra $100 might or might not help put a dent in that list, depends how far your willing to go with this.

I edit videos, render, ect. Must say video editing has been most expensive process yet. SSD gets limited by its 'boost' higher performance, not its consistive performance. Whereas 5xraid0 on wd 2tb hhd hits 500mbps consistant. Added 2 extra hhd for backup (raid 6) for the 5x in raid. But that said, when you render, your reading from/writing to/programs using. So if your doing all this to one chunk of hhd space, your getting that 500mbps, but splitting it 3 ways. So 500= 200 read/200write/100program. So to get the most, you need to have it all on separate drives, going to separate drives, with program on a separate drive. Normally if you don't 'optimize' your hhd/ssd to maintain a high transfer rate, it just takes longer to render.

Problem arises when your doing HD/4K (especially 4K this is essential). The transfer rate if not kept above a certain minimum (some light testing shows around 300-400mbps) and it takes longer to render a second than a second of actual movie, the fps of the movie itself gets fluctuated and looks horribly glitchy and just complete crap. 4k is huge data, high bitrate, high fps, with hardcore audio. Now as mentioned before, ssd do offer great read/write speeds (boost) but theyre consistency is horrible. Which is great for a bunch of small files, but prolonged large files are crap for ssd. The best way to maximize the rendering process is to have read speeds of 500-600mbps, write speeds of 500-600mbps, and audio on a separate drive (ssd's do work fine for this) around 500mbps. With os and rendering program on a separate drive. This requires 2 separate hhd raid assemblies at 4-6 hhd's a piece, an ssd for boot/programs, and a third ssd (or a third raid assembly which is just to much) for the audio.

Now I say all this because this is what ive found out. My first time doing this, I had my 8350, 500gb ssd, 3x3tb hhd's (not raid of any sort) and a 7870. Went to render first time, think I was reading/writing to same drive, and it was taking over the amount of time of the video, Checked task manager to see cpu at like 14%, gpu at like 4%, ram at 4%, but hhd at 100%. Now ive run a bunch of tests on this showing all the different settings and writing to/from ect different drives, showing all the %'s of components and which works the best. But to put it quickly, typicall hhd transfer rate 100mbps, ssd (consistency) around 300mbps. So if your reading and writing to same drive, your splitting that maximum transfer rate, plus (worse yet) the read fights the write and vice versa for more of that percentage. When you watch it, it goes from 50/50 split to 30/70 to 70/30 to 25/55 to 55/25 and all over, which is fine for lighter stuff, but like I said for 4k and of the sort, it turns it into an unplayable unwatchable mess of a movie.

But to actually put this into use on this forum. The diference between the amd and intel, depending what your dong, could easily be nothing. Depending what exactly your doing, which programs your using, what components you have, and how you have them set up, you could easily get by with a Pentium processor if your being limited by something else.
 
Another note, as far as price/performance, how willing you are to spend more, what means most to you, ect.
i7 4770k vs a i7 4930k is a huge difference. CPU wise, not huge huge but very noticeable. Your getting 6 core (12 with hyperthreading) and large cache increase. But its also a 2011 board setup, so you can take advantage of quad channel memory, as well as pci lanes, and often more sata ports with higher speeds. Plus the ability to run 2-4 GPS in sli/crossfire at full 16x because of build in plx chipsets.

And, on another/another/another note.....

GPU wise, depends on which programs your using and what they use. NVIDIA has cuda and AMD uses open cl and gl. I find that very most often programs work with both, but if not CUDA is the preffered/supported. Another thing to consider is that unlike gaming, which stays around 2gb ram utilized (other than skyrim modded and I belive titanfall which can use up to 3gb), most gpu accelerated programs can utilize most if not all your gpu ram. Which is important as its gddr5 vs ddr3, huge speed difference (normally at least 5000mhz compared to 1600/1866 that typically used in desktops. In the same price/performance gpu between NVidia and amd, amd typically gives more ram per card. But also amd typically runs warmer, with more power, and louder with fans than the NVidia counterpart.
 

Yea I already purchased the asus gtx 770
 


Hmm while considering that I already have purchased the graphics card it would probably just be better for me to get the i7 is what you are saying. It's really only a $50 difference because I would purchase a h100i to go with the 8350 and a hyper 212 evo for the 4770k which I could then upgrade cooling later on if need be
 


The graphics card was just one of the things you could spend the extra $$Dinero$$ on. But in all honesty you have the h100i and coolermaster worked out great (best coolers I would go with for either). The amd is only better at a very limited few things, is close enough to the intel to not care about the difference in several, but is surpassed by a small/sometimes significant margin by the intel. And for only $50 more, id opt for the intel i7. Plus you'll at the very very least save that $50 in the next three years in electricity (probably closer to just the first year).

On a side note, which motherboard do you have/plan on going with. The asus hero is by far the fan favorite (myself included). Otherwise the asus deluxe board is also very nice if your planning on adding some major hardware down the road.
 

I'm going with the hero motherboard. It also goes with my color scheme of black and red lol. My case is the nzxt h440 black/red

 
dblkk hit on something that's what I looked at too as to save here and put it there as in the i7 would be nice but did I realy need it? and then that extra cash between the i7 and 15 can go to a better gpu or board? its what you can do to find a nice balanced rig to fit your needs and do the tasks well.
 


If its straight gaming then i5 would be plenty. If theres any rendering or intensive programs at all the price difference is more than worth the i7. As for what else I would do or spend money on, since you already have gpu, maybe an ssd if you don't. Otherwise, it'd be cash for a video game, system wise hero board is top end. The formula is nice but doesn't add anything close to being worth $100 more. The extreme isn't even rated as well as the hero board, and that's $200 more or double the price. The msi boards all look nice (and at worst same price as the hero), but hero for what you get and spend is (my opinion) top notch. Z87 is pretty straight forward mobo wise, top end $200 hero, nothing better offers anything worth price. Z77 boards I wouldn't go with less than a asus rampage black for $500, but only because only reason id go z77 route is 4930k. Otherwise 4770k is best haswell in my eyes with hero being best board.
 
Solution


Not sure if trolling.

That is a horrible build. Much overpowered GPU for the CPU, the CPU is way too slow. Not enough RAM (no point for the really fast stuff) or disk space either.

OP already has the 770 and very likely doesn't need the software or the Blu ray burner.

A lot of 3d modelling applications are highly singlethread bound on manipulation tasks. Rendering however, uses all available CPU.

http://www.tomshardware.com/reviews/specviewperf-12-workstation-graphics-benchmark,3778-18.html

Not sure what the OP is using the maya and solidworks scale with clockspeed.
 
The R9 290 is likely twice as fast at the GTX 770 in video rendering via GPU acceleration, and multiple-times faster than using CPU cores.

Using fast RAMs in combination with a UNB OC can nearly double memory bandwidth and reduce latency. The AMD Athlon X4 750K Trinity will provide 90% of the gaming performance at a substantially reduced cost compared to any other CPUs, and the FM2+ platform provides an upgrade path to Richland and Kaveri, and possibly the next-gen Carizo.

That's why.

If the OP feels the overwhelming need to stick with the GTX770 and Intel CPU, he should likely wait on the 9-series Z/H motherboards and stick with something in the range of the 4570Ks.

 

Don't worry I'm not going to consider this guys build
 
PCPartPicker part list / Price breakdown by merchant / Benchmarks

CPU: Intel Xeon E3-1230 V3 3.3GHz Quad-Core Processor ($244.99 @ Amazon)
CPU Cooler: Cooler Master Hyper TX3 54.8 CFM Sleeve Bearing CPU Cooler ($18.16 @ NCIX US)
Motherboard: ASRock B85 Pro4 ATX LGA1150 Motherboard ($69.99 @ Amazon)
Memory: G.Skill Ares Series 8GB (2 x 4GB) DDR3-1600 Memory ($68.99 @ Newegg)
Storage: Crucial M500 240GB 2.5" Solid State Disk ($119.99 @ NCIX US)
Storage: Seagate Barracuda 3TB 3.5" 7200RPM Internal Hard Drive ($104.99 @ NCIX US)
Case: Corsair 200R ATX Mid Tower Case ($57.99 @ Micro Center)
Power Supply: XFX 650W 80+ Bronze Certified ATX Power Supply ($79.74 @ Amazon)
Optical Drive: Asus DRW-24F1ST DVD/CD Writer ($14.99 @ Newegg)
Total: $779.83
(Prices include shipping, taxes, and discounts when available.)
(Generated by PCPartPicker 2014-05-06 13:03 EDT-0400)
 


Haha, no that CPU would bottleneck the 290 in almost every game. Richland/trinity doesn't scale RAM past 2133 mhz due to an inefficient controller and thats mainly for the igp. The OP's software would also need to support GPU rendering.

The 750k is less than half the performance of the i7 4770k. With the 4770k the CPU is good for a few years as well and will easily support future GPU upgrades while the 750k will be horribly bottlenecked now and any GPU upgrade will need a CPU upgrade.

The OP already said he has a 770 as well.