Do I need to upgrade?

[Niege]

I've been a member a long time, but haven't posted in a few years owing to work commitments.

My rig is four years old, albeit with a bunch of upgrades over the years. My main use is for photo editing (Photoshop, Lightroom, etc.) followed by business applications. No gaming. I may go into video editing in a while as I incorporate that into my photo biz.

Current System:
Asus P6X85D-E mobo LGA 1366
i7-930 2.8Ghz
Win7 64 Home
16GB DDR3 12800 (could be 24GB if I install my upgrade copy of Win7 64 Pro - I don't want Win8)
Samsung 840 Pro SSD 256GB
3 additional internal HDs for storage and internal backup. Several external HDs.
EVGA Nvidia Geforce GTS250 1GB DDR3 2200Mhz
Seasonic S12 650W power supply

I'm not sure if a performance gain would be worth it to swap out mobo and cpu. GPU is fine for what I do now but not sure that would be okay for video editing.

Any insights would be great. Thanks!

 

[Cerb]

No need whatsoever. A new Haswell CPU would be faster, but you wouldn't see a doubling of performance in any use case.

 

[aigomorla]

aftermarket heat sink + overclocking will extend the life of your machine and may make it feel a bit faster.

You can easily net 800mhz+ on that processor with a good after market heat sink.

 

[Zorander]

I second overclocking the CPU. There is quite a bit of overclocking headroom there.

I might be wrong but I believe Windows 7 Home does not support more than 16Gb?

 

[MongGrel]

Xeon L5639 Overclocking on X58

http://forums.anandtech.com/showthread.php?t=2335636&page=56&highlight=l5639


Not necessary, but still, if you're doing video.

 

[crashtech]

Seems like upgrading is a bear market these days, but who has put any raw numbers out there to support the lack of need? Even compared to an i7-965 at 3.2GHz, the 4770K wins by large margins in just about everything:

http://www.anandtech.com/bench/product/45?vs=836

So while there is no "need," upgrading certainly will result in a very noticeable performance increase without the worry of overclocking a business machine. I'd actually suggest a 4790, but there is little data on that one in the Bench. It would score slightly better than a stock speed 4770K, of course.

 

[NewYorksFinest]

A 2 letter word that begins with an N and ends with an O :biggrin:.

 

[Cerb]

A 2 letter word that begins with an N and ends with an O :biggrin:.

True.

But, as mentioned, overclocking would help. Those CPUs tended to be good for 3.5GHz+ w/o much voltage, and yet a few hundred MHz higher than that, depending on how much time you wanted to sink into testing OCs.

Also, most of what the new CPUs do so much better in benchmarks is loops in benchmarks that can use AVX2, and/or benchmarks with GPU use mixed in, so take some of those with more than a few grains of salt.

 

[NewYorksFinest]

True.

But, as mentioned, overclocking would help. Those CPUs tended to be good for 3.5GHz+ w/o much voltage, and yet a few hundred MHz higher than that, depending on how much time you wanted to sink into testing OCs.

Also, most of what the new CPUs do so much better in benchmarks is loops in benchmarks that can use AVX2, and/or benchmarks with GPU use mixed in, so take some of those with more than a few grains of salt.

The i7-930 is just fine. Maybe upgrade the GPU to a GTX 750 Ti for $130 or something, but the i7-930 is a fine CPU. JMHO.

 

[crashtech]

Keep that salt shaker handy, the 4770 finished applying filters twice as fast as an i7-860 in this benchmark:

http://www.tomshardware.com/charts/cpu-charts-2013/-33-Adobe-Photoshop-CS-6,3169.html

 

[Niege]

Wow. Thanks, guys. I'm running it stock right now but I do have a SCYTHE, Mugen-2 Rev.B (SCMG-2100) CPU Cooler. Big honking thing. I can actually run without a CPU fan with just the radiant fins on this thing, though definitely not a good idea.

I haven't tried OC yet, valuing stability over bleeding edge performance. Definitely food for thought here. Thanks, again.

 

[Cerb]

^ Where's a link to the benchmark setup? ~50% faster than a i7-3770 is going to need GPU acceleration, and/or AVX2. The benefits of both vary greatly depending on the test setup and filters used (also, CC seems to have at least some filters that can justify getting much better than IGP).

Not saying the new chip won't be a good bit faster, but this is a case where blindly trusting those benchmarks can lead to disappointment after dropping $500+ on new gear, and not getting the results those benchmarks would indicate, since the improvements are very much based on what you do as a user (which I'm sure is one of several reasons Adobe does not have a public Photoshop benchmarking tool, though I'm sure they have dozens of internal ones).

 

[schmuckley]

You should OC more

 

[crashtech]

Either we trust benchmarks, anecdotal evidence, or nothing at all. While I agree with the grains of salt premise, I'm not ready to say benchmarks are totally useless. They're really all we have to hang onto in a sea of misinformation.

 

[Torn Mind]

Benchmarks are arguably the best of the worst ways to discern "tasks done per clock", which is the companion variable to clockspeed. These two things determine has fast CPUs get things done(things being "operations").

A Haswell i7 is likely to be at least 1.33 times faster than your i7 at stock, depending on what is being benched.

http://anandtech.com/bench/product/836?vs=46
Divide the larger number by the smaller one. Make sure to check whether lower is better or higher is better when looking at these things.

With the release of Devil's Canyon this month, I would advise you to get the 4790K when you feel like picking up video editing in the near future. It has a 4.0 Ghz base clock and 4.4 Ghz Turbo Boost. For a fellow who values stability, it should be a worthy upgrade. Or if video editing is to be incorporated, say, Q2 of next year, you can get Broadwell-K instead.

Every frame counts in video editing, and the faster you got through them, the more time you have to do other things.

 

[Charlie98]

I haven't tried OC yet, valuing stability over bleeding edge performance. Definitely food for thought here. Thanks, again.

Just remember... building an OC doesn't mean you have to sacrifice stability for performance. Just don't go crazy with it. My business machine has an OC'ed i5, pretty much as reliable as the day is long.

 

[Cerb]

Either we trust benchmarks, anecdotal evidence, or nothing at all. While I agree with the grains of salt premise, I'm not ready to say benchmarks are totally useless. They're really all we have to hang onto in a sea of misinformation.

Benchmarks aren't useless. Benchmark scores of recent Adobe Photoshop, like CS 6 or CC, that are just single scores, without clear configuration and test details, are on the order of usefulness of Passmark scores.

While not useful for current hardware (or software, if the OP is using CC), this review covered all the bases, and then some, for example.