HardOCP compare a 2600K with a 7700K

[Ayahuasca]

2500k to a 7700k is understandable it's the ones who upgrade every gen I don't understand, you're pretty much getting nothing for your money that your eyes will notice or brain perceive.

 

[Hedge]

The main justification for me is the fact that sandybridge was so long ago that I actually didn't mind throwing a bit money at my pc.

Thats good enough for most!

 

[DragonQ]

The main justification for me is the fact that sandybridge was so long ago that I actually didn't mind throwing a bit money at my pc.

Thats good enough for most!

I believe around these parts that's called the upgrade itch.

I sort of have it myself but step 1 is buying a better monitor. Step 2 is GPU, and step 3, if needed, will be the rest. With FreeSync 2 just being announced I'm now hesitant to even go for step 1, let alone step 3.

 

[Hedge]

I agree entirely with your logic. I've been holding off on a monitor for ages because I feel they've offered nothing to justify their huge price.

Then unfortunately now that a noticeable monitor advancement is on the horizon pcs aren't nearly up to the task of making the most of them.

 

[Sampsy]

Showing maximum results at 640x480 is about as pointless as it gets for people wanting real answers.

I've never quite understood the point of benchmarks like those and HARDOCP are hardly the only guilty party. Benching games at resolutions nobody would use, running K skew processors at stock or matching clockspeeds. They are informative if you're interested in investigating how two CPUs differ with regard to some metric such as IPC but ultimately isn't particularly relevant to consumers. The best benchmarks are those which closely mimic the real-world usage scenario consumers have in mind and are performed using a realistic hardware setup.

That means don't bench Metro at 768p - bench something like BF1 at 1080p, 1440p & 4k in a full 64 player game. Why not bench ARMA or something like GW2 - games which still have active playerbases which could also greatly benefit from a boost in CPU performance. Throw in some popular CPU-friendly games and some which place an intermediate load on the CPU to be representative. Further when benching CPUs you need more information than the average FPS. An averge FPS of 200 can play significantly worse than an average FPS of 40 if every 3 seconds the 200FPS system stutters. Bench some popular professional applications too by performing a realistic task. However, don’t bother with synthetic benchmarks as I couldn’t care less how many imaginary performance points my PC gets – not when I’m thinking about spending money on an upgrade anyway.

There’s also little point benching an overclockable CPU at stock as nobody should be running it at stock. Similarly, there’s no point matching the clock speed to some other CPU because no consumer would OC their CPU like that. If I have a 2600k in my PC I’m going to overclock that as much as I reasonably can and if I bought a 7700k I’d do the same. Maybe one is a better overclocker than the other and that will impact my purchasing decision. However, if a review benches them at the same speed I don’t actually learn anything relevant to my purchasing decision. In similar logic don’t try to match the memory speed either. I’m not going to spend all that money on new RAM just so I can have as small an upgrade as possible. I’ll buy the best RAM I can which fits in my budget. It helps to look at the impact the CPU alone has - but I wouldn't ever just buy a new CPU - I'd buy a new CPU, Motherboard and RAM at the very least.

Lots of these reviews try so hard to be scientific they end up deviating so far from real-life use cases and fail to use setups consumers either likely have or would realistically consider upgrading to that they are no longer useful.

 

[Silent_Scone]

I've never quite understood the point of benchmarks like those and HARDOCP are hardly the only guilty party. Benching games at resolutions nobody would use, running K skew processors at stock or matching clockspeeds. They are informative if you're interested in investigating how two CPUs differ with regard to some metric such as IPC but ultimately isn't particularly relevant to consumers. The best benchmarks are those which closely mimic the real-world usage scenario consumers have in mind and are performed using a realistic hardware setup.

That means don't bench Metro at 768p - bench something like BF1 at 1080p, 1440p & 4k in a full 64 player game. Why not bench ARMA or something like GW2 - games which still have active playerbases which could also greatly benefit from a boost in CPU performance. Throw in some popular CPU-friendly games and some which place an intermediate load on the CPU to be representative. Further when benching CPUs you need more information than the average FPS. An averge FPS of 200 can play significantly worse than an average FPS of 40 if every 3 seconds the 200FPS system stutters. Bench some popular professional applications too by performing a realistic task. However, don’t bother with synthetic benchmarks as I couldn’t care less how many imaginary performance points my PC gets – not when I’m thinking about spending money on an upgrade anyway.

There’s also little point benching an overclockable CPU at stock as nobody should be running it at stock. Similarly, there’s no point matching the clock speed to some other CPU because no consumer would OC their CPU like that. If I have a 2600k in my PC I’m going to overclock that as much as I reasonably can and if I bought a 7700k I’d do the same. Maybe one is a better overclocker than the other and that will impact my purchasing decision. However, if a review benches them at the same speed I don’t actually learn anything relevant to my purchasing decision. In similar logic don’t try to match the memory speed either. I’m not going to spend all that money on new RAM just so I can have as small an upgrade as possible. I’ll buy the best RAM I can which fits in my budget. It helps to look at the impact the CPU alone has - but I wouldn't ever just buy a new CPU - I'd buy a new CPU, Motherboard and RAM at the very least.

Lots of these reviews try so hard to be scientific they end up deviating so far from real-life use cases and fail to use setups consumers either likely have or would realistically consider upgrading to that they are no longer useful.

I wouldn't even call it scientific, I'd call it misguided

 

[Kaapstad]

I've never quite understood the point of benchmarks like those and HARDOCP are hardly the only guilty party. Benching games at resolutions nobody would use, running K skew processors at stock or matching clockspeeds. They are informative if you're interested in investigating how two CPUs differ with regard to some metric such as IPC but ultimately isn't particularly relevant to consumers. The best benchmarks are those which closely mimic the real-world usage scenario consumers have in mind and are performed using a realistic hardware setup.

That means don't bench Metro at 768p - bench something like BF1 at 1080p, 1440p & 4k in a full 64 player game. Why not bench ARMA or something like GW2 - games which still have active playerbases which could also greatly benefit from a boost in CPU performance. Throw in some popular CPU-friendly games and some which place an intermediate load on the CPU to be representative. Further when benching CPUs you need more information than the average FPS. An averge FPS of 200 can play significantly worse than an average FPS of 40 if every 3 seconds the 200FPS system stutters. Bench some popular professional applications too by performing a realistic task. However, don’t bother with synthetic benchmarks as I couldn’t care less how many imaginary performance points my PC gets – not when I’m thinking about spending money on an upgrade anyway.

There’s also little point benching an overclockable CPU at stock as nobody should be running it at stock. Similarly, there’s no point matching the clock speed to some other CPU because no consumer would OC their CPU like that. If I have a 2600k in my PC I’m going to overclock that as much as I reasonably can and if I bought a 7700k I’d do the same. Maybe one is a better overclocker than the other and that will impact my purchasing decision. However, if a review benches them at the same speed I don’t actually learn anything relevant to my purchasing decision. In similar logic don’t try to match the memory speed either. I’m not going to spend all that money on new RAM just so I can have as small an upgrade as possible. I’ll buy the best RAM I can which fits in my budget. It helps to look at the impact the CPU alone has - but I wouldn't ever just buy a new CPU - I'd buy a new CPU, Motherboard and RAM at the very least.

Lots of these reviews try so hard to be scientific they end up deviating so far from real-life use cases and fail to use setups consumers either likely have or would realistically consider upgrading to that they are no longer useful.

Nothing scientific about HARDOCP, if they used 1440p and 2160p with decent game settings the results between CPUs would be tiny. The problem with HARDOCP doing this is they would have to find something else to write about this week for clickbait.

 

[Uriel]

It's interesting. In HardOCP's GPU reviews they usually test at highest playable settings and report on that basis. I find it very interesting to see what visual effects can be enabled on different cards at a given framerate rather than to see what the framerate is at the same settings.

This CPU comparison seems like quite the opposite approach.

 

[pingu666]

isnt some of the performance gain from new intels from the memory?

and measuring miniums and frametimes has come about mostly from stagnation in performance

 

[Silent_Scone]

isnt some of the performance gain from new intels from the memory?

and measuring miniums and frametimes has come about mostly from stagnation in performance

Yes, memory helps. Will increase the gap a fair bit in certain instances. This is why this sort of testing is futile. If you ran at 1080p and 1440p with the 7700K running 3200Mhz on the DRAM, the gap would be bigger. That includes minimums, which over the last year or so journos have seemingly forgotten about.

 

[Ayahuasca]

I'm tempted to loosen the timings on my 2400mhz 10-12-12-28-1T DDR4 kit and see if higher mhz helps.

 

[Silent_Scone]

I'm tempted to loosen the timings on my 2400mhz 10-12-12-28-1T DDR4 kit and see if higher mhz helps.

Probably not worth pursuing that. 3000-3600 is where you'll see a notable difference in varying frame rate

 

[Dave2150]

I don't know why anyone is currently interested/fussed over Kabylake - it's such a better option to wait a few weeks for Ryzen, as it will have a big effect on prices, as well as most likely outperforming all Kabylake offerings.

 

[Buckster]

I can't either Dave - I'd certainly not even consider upgrading my 3770k to one - would quite literally be a waste of money

I thought the HardOCP article was fair - I'm sure I read somewhere their explanation of why used low resolutions

what they are showing is even with the "ideal" scenarios - theres very little real difference

come on AMD

 

[SoundsGood]

So I've just gone from 2700K @ 4.6Ghz to 6600 (non K). I did some benchmarks before and after. Hopefully interesting to someone

Battlefield 1 @ Ultra 1440p - fps taken stood still at the same part of the campaign
2700K @ 4.6GHz: 56fps
6600 @ stock: 85fps
6600 @ 4.4GHz (via BCLK): 85fps

3dMark Basic Edition
2700K @ 4.6GHz: 5568
6600 @ stock: 5548
6600 @ 4.4GHz (via BCLK): 5699

2700K @ 4.6GHz: http://www.userbenchmark.com/UserRun/2563476
6600 @ stock: http://www.userbenchmark.com/UserRun/2569399
6600 @ 4.4GHz (via BCLK): http://www.userbenchmark.com/UserRun/2570746

 

[paul10]

If your rendering you may do better with 6 cores on x99 especially if you overclock.

maybe but my software might not benefit so much i use edius workshop and the new chip supports h265 and intel quick sync which the 6 cores chips do not ?

so maybe a overclocked 7700k will be better than 6 core overclocked doubt little difference in cost though ,i am sure it will be better than my i7 950 whatever the outcome

 

[~>Dg<~]

So I've just gone from 2700K @ 4.6Ghz to 6600 (non K). I did some benchmarks before and after. Hopefully interesting to someone

Battlefield 1 @ Ultra 1440p - fps taken stood still at the same part of the campaign
2700K @ 4.6GHz: 56fps
6600 @ stock: 85fps
6600 @ 4.4GHz (via BCLK): 85fps

3dMark Basic Edition
2700K @ 4.6GHz: 5568
6600 @ stock: 5548
6600 @ 4.4GHz (via BCLK): 5699

2700K @ 4.6GHz: http://www.userbenchmark.com/UserRun/2563476
6600 @ stock: http://www.userbenchmark.com/UserRun/2569399
6600 @ 4.4GHz (via BCLK): http://www.userbenchmark.com/UserRun/2570746

as great as your effort is you doing the campaign also stood still.

for real benchmarks.go into a full 64 man game scar for eg and go into action and play as normal.not looking at sky or floor.this gives you higher fps .

just play as normal 1 minute runs in action ( i do 3 to avg it out and do min max and avg ).you have to do mp because campaign is so different.hardly any difference between many cpus.

 

[Silent_Scone]

I don't know why anyone is currently interested/fussed over Kabylake - it's such a better option to wait a few weeks for Ryzen, as it will have a big effect on prices, as well as most likely outperforming all Kabylake offerings.

That's a really interesting comment. I thought only last week you were saying how games don't benefit from multicore execution?

Considering Ryzen will at best match Broadwell IPC, how are you expecting it to outperform Kaby?

lol.

 

[Minstadave]

That's a really interesting comment. I thought only last week you were saying how games don't benefit from multicore execution?

Considering Ryzen will at best match Broadwell IPC, how are you expecting it to outperform Kaby?

lol.

I'm sure he's the same chap that argued blue in the face that Skylake was better than anything X99 at Skylake's release because of the better IPC.

 

[8 Pack]

Again and again this stuff. Just wait and see about Ryzen type nothing till you know.....