£300 to spend...
Page 3 of 5 Goto page Previous  1, 2, 3, 4, 5  Next
Morphineus
VIP Member



Posts: 24883
Location: Sweden
PostPosted: Tue, 6th Oct 2015 02:16    Post subject:
Janz wrote:
but that should only happen if the board is in the store shelf since over 6 months


Happened to my girl's Mobo.

But it was good, they said it was compatible for all i7 processors (didn't specify on haswell refresh) on their store page. So I abused that fact. Got a z97 from a higher price range (and better model) as compensation. Razz

Could have easily updated it, but I saw an opportunity and went for it. Twisted Evil

Tho, twas my girl that made me aware of their miss information. I knew when we ordered that we had to do it. She just didn't agree with it Razz


Back to top
sabin1981
Mostly Cursed



Posts: 87805

PostPosted: Tue, 6th Oct 2015 21:50    Post subject:
sabin1981 wrote:
As a SB owner, I can also confirm that the proc is simply getting too long in the tooth now. It simply isn't as capable any more and my astounding amount of personal experience with an equally astounding amount of games confirms that. I'd love to snag a Haswell setup but the price is simply too prohibitive, so for the foreseeable future my only option *at all* would be to stick with SB but to upgrade to an i7 instead to at least make use of multithreading in the games that benefit from it. Crysis 3 was a particularly unpleasant experience where I was completely CPU restricted right from the start.

Had I the choice? £300 in my pocket? No contest, Haswell board + Haswell i7 if I could wing it.


Further confirmation when playing Vermintide, the game struggles to maintain good framerates in combat - often falling down to high 40s, low 50s, with CPU usage pegging 80-99% across all four cores. The game doesn't look anything special. SB procs, at least quads without HT, are simply too old.. even if they are still great.
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Tue, 6th Oct 2015 22:12    Post subject:
yeah, mp games are often more CPU hungry. battlefield 4 is a good example of that. sp and mp totally different cpu usage.


steam: Drowning witch
uplay: Haxeety
game completion log:
http://pastebin.com/g6MgD5DV
rig: http://www.userbenchmark.com/UserRun/2090866
Back to top
Breezer_




Posts: 10833
Location: Finland
PostPosted: Tue, 6th Oct 2015 22:35    Post subject:
i5 in general is not really up for the task these days on some ports, games seem to use i7 nicely.
Back to top
MinderMast




Posts: 6172

PostPosted: Tue, 6th Oct 2015 23:10    Post subject:
There aren't that many games where HT makes a lot of difference yet and it's definitely not enough on its own. I don't think an SB i7 is gonna be any good, for example, since in my experience the 4790K with HT off did a lot better than the 2500K at pretty much the same clocks. Meaning it was mainly the CPU architecture that helped.

I guess the reality is that you need both: at least a Haswell generation CPU and preferably an i7 at that.
Back to top
VGAdeadcafe




Posts: 22230
Location: ★ ಠ_ಠ ★
PostPosted: Tue, 6th Oct 2015 23:18    Post subject:
Hyperthreading is only really needed in dual cores.
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Tue, 6th Oct 2015 23:27    Post subject:
few months old article on the topic:

http://www.techspot.com/review/972-intel-core-i3-vs-i5-vs-i7/


Quote:
Gaming is a different story. As we've seen time and time again when benchmarking PC games, the Core i7 is generally overkill as games that utilize the extra threads are relatively rare. The best example we have at hand is Metro: Redux which was 8% faster at 1080p using a Core i7 over a Core i5 with the GeForce GTX 980.
When paired with the GTX 980 (a very fast GPU by all means), the i7-4790 was 1% faster than the i5-4690 at 1680x1050 and no faster at 1080p. When accompanied by the more modest GTX 960, the Core i7-4790 was on par with the Core i5-4690 at 1680x1050 and 1% slower at 1080p.



and a dude on guru3d did his own benches on some other games

http://forums.guru3d.com/showthread.php?t=402869

another user batch of tests:

http://www.overclock.net/a/intel-core-i3-vs-core-i5-vs-core-i7-gaming-performance-with-geforce-gtx970
Back to top
MinderMast




Posts: 6172

PostPosted: Tue, 6th Oct 2015 23:35    Post subject:
One of the recent games that named an i7 as a recommended spec was Witcher 3...
I checked out this analysis:


Note how at around 1:30 mark, when he rides into the city the 4670K starts to lag behind the 4790K. It's apparent from the framerate, but the frametimes on the right also show the i5 struggle quite a bit the longer it goes on.
With a CPU causing such performance issues, it won't feel smooth even with a relatively high framerate.
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Tue, 6th Oct 2015 23:53    Post subject:
yep, if you are very sensitive to frame times, i7 is the best. I personally do not notice any lack of smoothness in witcher 3 with a 4.6ghz 4670k.


steam: Drowning witch
uplay: Haxeety
game completion log:
http://pastebin.com/g6MgD5DV
rig: http://www.userbenchmark.com/UserRun/2090866
Back to top
StrEagle




Posts: 14059
Location: Balkans
PostPosted: Tue, 6th Oct 2015 23:59    Post subject:
OMFG 1080p TITAN X 80FPS Shocked
I need to upgrade my 970 Laughing


Lutzifer wrote:
and yes, mine is only average
Back to top
scaramonga




Posts: 9800

PostPosted: Wed, 7th Oct 2015 00:18    Post subject:
Drowning_witch wrote:


and a dude on guru3d did his own benches on some other games



Laughing Laughing
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Wed, 7th Oct 2015 00:55    Post subject:
@scaramonga, i hesitated posting that link indeed Laughing

MinderMast wrote:

Note how at around 1:30 mark, when he rides into the city the 4670K starts to lag behind the 4790K. It's apparent from the framerate, but the frametimes on the right also show the i5 struggle quite a bit the longer it goes on.
With a CPU causing such performance issues, it won't feel smooth even with a relatively high framerate.



It's either the patches, or the 4.6ghz overclock on my i5 4670k, but I just downloaded a save file in novigrad, and spent a few min on horseback in the city, frametime was constantly between 16-18ms, even in super crowded areas. i doubt anyone is that sensitive, but you never know Very Happy
Back to top
StrEagle




Posts: 14059
Location: Balkans
PostPosted: Wed, 7th Oct 2015 00:58    Post subject:
if you want better performance at multiplayer games, get a dedicated LAN card Rolling Eyes


Lutzifer wrote:
and yes, mine is only average
Back to top
MinderMast




Posts: 6172

PostPosted: Wed, 7th Oct 2015 07:14    Post subject:
Drowning_witch wrote:
@scaramonga, i hesitated posting that link indeed Laughing

MinderMast wrote:

Note how at around 1:30 mark, when he rides into the city the 4670K starts to lag behind the 4790K. It's apparent from the framerate, but the frametimes on the right also show the i5 struggle quite a bit the longer it goes on.
With a CPU causing such performance issues, it won't feel smooth even with a relatively high framerate.



It's either the patches, or the 4.6ghz overclock on my i5 4670k, but I just downloaded a save file in novigrad, and spent a few min on horseback in the city, frametime was constantly between 16-18ms, even in super crowded areas. i doubt anyone is that sensitive, but you never know Very Happy

For some reason I was always sensitive to such things, and it only got worse when I got a 120Hz monitor Sad
That benchmark also shows the framerate staying above 60FPS even when the CPU starts to bottleneck, so you might just get away with it if that is your cap Smile

Still, the main point is that it looks clear enough that an i7 has the advantage (or it had, at some point) when it comes to CPU intensive tasks in this game. If you pay decent money for a high end GPU you probably wouldn't want it to be held back by your CPU at any point.

In general, it doesn't really help that almost every benchmark I have seen trying to show that difference (or lack of it) goes about it the wrong way. It was quite annoying when I was considering my upgrade choices between 2500K to 3770K or getting a new system with a 4790K. Even knowing the limitations of the 2500K form personal experience, I couldn't find any reliable benchmark that would show them, so I couldn't trust any conclusions made on the higher-end CPUs.
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Wed, 7th Oct 2015 16:24    Post subject:
yeah, if you have the budget, it's a no brainer to go for i7, there are games where the differences are notable, and if you are going for above 60 fps, it's pretty much a must to get the best CPU to keep up with the gpu as much as possible.

I really like those digitalfoundry/eurogamer benches. atm, I think they are the best representation of gpu and cpu benches you can get from a third party source atm.

i've even compared some of their vids on my rig, and they were pretty much identical.
Back to top
sabin1981
Mostly Cursed



Posts: 87805

PostPosted: Wed, 7th Oct 2015 17:01    Post subject:
What I'd like is to see the same test ran, but this time with an i5-2500k and i7-2600k versus the 4690k and 4790k. I'm sure there's a fairly substantial difference, 10fps easily, and that can mean the difference between barely sustaining 60 (as that video shows) and barely sustaining 50 (as my experience was)
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Wed, 7th Oct 2015 19:33    Post subject:
here is a 3 month newer bench with witcher 3 again, this one does include the 2500k. looks like patches made it less cpu bound. but gta V and crysis 3 min fps is still killing the 2500k a bit, not to mention they show a very undemanding scene in crysis 3.

Back to top
sabin1981
Mostly Cursed



Posts: 87805

PostPosted: Wed, 7th Oct 2015 20:04    Post subject:
Right across the board there's a clear 15-20fps difference between the 6600k and 2500k.. that's insane Shocked But yup, minimum is what matters and that is where the 25 just gets it ass kicked by strangely CPU-heavy scenes. Crysis 3 was a goddamn mess for me whenever foliage was on screen, same with GTAV.. I had to run with "normal" grass in that because the perf drop was terrible.
Back to top
MinderMast




Posts: 6172

PostPosted: Wed, 7th Oct 2015 20:20    Post subject:
It's hard to say because this benchmark seems to have the CPUs overclocked, while the previous one didn't (or at least they didn't specifically list them as OC). Performance for the 4690K is only slightly better in some cases compared with the previous benchmark and still behind the 4790K (supposedly at stock clocks). [EDIT: Uhh, got it wrong, it was the GPU that was overclocked, not the CPU].

It is, however, surprising to see Skylake suddenly surging ahead with SB, IB and Haswell performing the same for the most part. The 6600K seems to give performance comparable to the 4790K in the previous benchmark.

Here are benches that are more relevant to the i5 vs i7 debate though:


i7 CPUs are doing a lot better all around. 2500K vs 2700K seems to give a 20 FPS boost here, if you compare the two benchmarks.

Another interesting thing is here:


Looks like with the i7 6700K you hit the GPU bottleneck and overclocking it doesn't make any difference. The 6600K lags behind on stock but catches up with the i7 when overclocked.

Still weird seeing Skylake leaping ahead while leaving the previous gen on the same level pretty much. I recall Haswell being in the same spot before, and now it looks like it's not much better than SB...

EDIT: I suppose I should mention that I mainly based this on W3 performance. In other ones it seems like a 6700K is the only real choice. Dat frame-time consistency... everything else plain sucks in comparison Very Happy


Last edited by MinderMast on Wed, 7th Oct 2015 20:56; edited 3 times in total
Back to top
couleur
[Moderator] Janitor



Posts: 14382

PostPosted: Wed, 7th Oct 2015 20:24    Post subject:
Drowning_witch wrote:
here is a 3 month newer bench with witcher 3 again, this one does include the 2500k. looks like patches made it less cpu bound. but gta V and crysis 3 min fps is still killing the 2500k a bit, not to mention they show a very undemanding scene in crysis 3.



How are thes procs clocked for that comparison? Same clocks?

It was always clear that at every new iteration the intel procs get a few % more performance. So having 10-30% more performance between 2500K and 6600K should be pretty normal.

On the other hand, those old 2500K still perform pretty good, considering you can also easily overclock them to 20-25% of their original clocks.


"Enlightenment is man's emergence from his self-imposed nonage. Nonage is the inability to use one's own understanding without another's guidance. This nonage is self-imposed if its cause lies not in lack of understanding but in indecision and lack of courage to use one's own mind without another's guidance. Dare to know! (Sapere aude.) "Have the courage to use your own understanding," is therefore the motto of the enlightenment."
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Wed, 7th Oct 2015 21:25    Post subject:
unless specified, it's stock clocks.


steam: Drowning witch
uplay: Haxeety
game completion log:
http://pastebin.com/g6MgD5DV
rig: http://www.userbenchmark.com/UserRun/2090866
Back to top
Breezer_




Posts: 10833
Location: Finland
PostPosted: Wed, 7th Oct 2015 21:53    Post subject:
Its funny that you guys now found out that HT is not that bad afterall.
Back to top
sabin1981
Mostly Cursed



Posts: 87805

PostPosted: Wed, 7th Oct 2015 23:39    Post subject:
Breezer_ wrote:
Its funny that you guys now found out that HT is not that bad afterall.


To be fair, it's taken this long to actually get a decent amount of games that utilise it Troll Dad
Back to top
Janz




Posts: 14001

PostPosted: Wed, 7th Oct 2015 23:53    Post subject:
you mean crappy ps4/xboner ports? Twisted Evil
Back to top
Breezer_




Posts: 10833
Location: Finland
PostPosted: Thu, 8th Oct 2015 08:27    Post subject:
sabin1981 wrote:
Breezer_ wrote:
Its funny that you guys now found out that HT is not that bad afterall.


To be fair, it's taken this long to actually get a decent amount of games that utilise it Troll Dad


We can only thank PS4 and Xbone calculator AMD APU "8-core" chips for it, devs need to squeeze everything from the CPU, so engines are getting heavily multithreaded.
Back to top
russ80




Posts: 4679
Location: Romania
PostPosted: Thu, 8th Oct 2015 15:26    Post subject:
Yeah love the part where people were saying HT / i7 doesn't make a difference in games and everyone should just buy the i5. Then surprise!



Main PC : I7 12700, MSI Ventus RTX 4090 24gb, Alienware AW3423DW QD-OLED
Laptop : I5 4200H @ 3400mhz boost, GTX 850m 2gb Vram DDR3, 4gb RAM DDR3
Derpsole : Playstation 5 disc edition, Ninty Switcherino
TV+audio: LG CX 65" / Sonos ARC + SL ones + Sonos sub 3
VR Headset: Meta quest 2 airlinked
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Thu, 8th Oct 2015 18:37    Post subject:
Anyone with an i7, can you test this area in witcher 3?


it's a forest area, and i think the ultra shadow settings might be cpu bottlenecked, as I gain 20 fps by lowering ultra shadows to high, but changing resolution with ultra shadows doesn't increase fps at all (it does with other shadow settings)

I mean over 20fps increase on a gtx970 with a notch down in 1 setting. that's insane Surprised




Back to top
Breezer_




Posts: 10833
Location: Finland
PostPosted: Thu, 8th Oct 2015 19:26    Post subject:
Dont think you are CPU bottlenecked there, since on both cases CPU usage is only little over 50%.
Back to top
couleur
[Moderator] Janitor



Posts: 14382

PostPosted: Thu, 8th Oct 2015 19:32    Post subject:
russ80 wrote:
Yeah love the part where people were saying HT / i7 doesn't make a difference in games and everyone should just buy the i5. Then surprise!




In 2006 dual core processors were irrelevant, now we use more than 4 cores.

Surprise. lol wut


Of course it was known that at some pointe it would become useful to have more than 4 cores. No surprise there. And 4 cores are still more than enough in most cases.


"Enlightenment is man's emergence from his self-imposed nonage. Nonage is the inability to use one's own understanding without another's guidance. This nonage is self-imposed if its cause lies not in lack of understanding but in indecision and lack of courage to use one's own mind without another's guidance. Dare to know! (Sapere aude.) "Have the courage to use your own understanding," is therefore the motto of the enlightenment."
Back to top
Drowning_witch




Posts: 10818
Location: Strawberry fields
PostPosted: Thu, 8th Oct 2015 19:35    Post subject:
Breezer_ wrote:
Dont think you are CPU bottlenecked there, since on both cases CPU usage is only little over 50%.


yeah, but some games do that, especially with nvidia driver multi threading.

for example, I recently replayed crysis 1, and had cpu bottlenecking since that engine is unoptimized.

with multithreading on core usage was 30% each, without it, one core was at 100% the others at 10%. but same fps in certain limited scenarios.

Just seems insane that a shadow setting drop gains 22 fps, and only 5 fps increase by going to 1024x768 with ultra shadows.

it would just be interesting to see if anyone else gets such huge fps increase with the shadow setting.

even have an autosave at the same exact location, view angle Very Happy

http://www28.zippyshare.com/v/lA2RltJH/file.html
Back to top
Page 3 of 5 All times are GMT + 1 Hour
NFOHump.com Forum Index - Hardware Zone Goto page Previous  1, 2, 3, 4, 5  Next
Signature/Avatar nuking: none (can be changed in your profile)  


Display posts from previous:   

Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum


Powered by phpBB 2.0.8 © 2001, 2002 phpBB Group