94 replies to this topic

[darqsyde]

Catamount, on 05 December 2014 - 07:48 AM, said:

There are numerous situations where an i5 at 4.5 is going to match or overtake an i7 at 4 and that's assuming you can get 4ghz on stock cooling (Ivy could, does Haswell?).

Haswell-Refresh i7-4790k comes at 4, turbo to 4.4, stock.

Edited by darqsyde, 21 December 2014 - 10:52 PM.

[Catamount]

darqsyde, on 21 December 2014 - 10:51 PM, said:

Haswell-Refresh i7-4790k comes at 4, turbo to 4.4, stock.

Toms' early testing of actual retail samples of Haswell showed absolutely abysmal overclocking, to the point that they were almost expressing anger at Intel for sending out review samples that were golden chips at OCing while none of the retail samples they were buying were coming anywhere remotely close to them. The article did not have a tone that came across as impressed with Intel.

It seems, however, that as time has gone on, Haswell chips have gotten better, to the point that I no longer point overclockers to Ivy. When I made the comment you replied to, I was not aware that Intel had sorted out the OC issues. It is worth noting that although Devil's Canyon does come stock at higher clocks, testing shows it doesn't reach higher OCs than normal Haswell chips (higher stock, less headroom, same end result), so for OCers I'd recommend the 4770k over the 4790k. Also keep in mind that that 4.4 is single core turbo. That's why Intel lists that as the "max turbo" frequency. The highest four-core turbo is going to be 4.1 or 4.2. I'd guess 4.2, but Intel ARK doesn't publish that stat, so that's just a guess based on observation of other chips.

Edited by Catamount, 22 December 2014 - 07:47 AM.

[Revener]

Game runs just fine on my AMD 8350 with a GTX 780.

and it ran fine on my old Phenom II X4 940 with GTX 275.

but yeah intel is better when you want single core performance .

Edited by Revener, 22 December 2014 - 05:22 PM.

[Bill Lumbar]

Revener, on 22 December 2014 - 08:16 AM, said:

Game runs just fine on my AMD 8350 with a GTX 780.

and it ran fine on my old Phenom II X4 940 with GTF 275.

but yeah intel is better when you want single core performance .

Same here for my FX-8350 with a XFX 7970 3gb. I also used to own the Phenom II X4 940 B.E. I never played this game with it, but it was a very solid CPU indeed. Given the fact that it kept pace with my Intel Q9550 even at 3.8 ghz OC that I still have, (a year and a half ago traded my 940 B.E. in a mid end gaming rig for a 2010 IMac 24" with a guy that wanted to try out a windows gaming rig.) I got my money back and (I sold the Imac 24" for $700 cash) use out of it for a solid 4 years.

[Thorqemada]

My FX8350 + R9 280x + Win7 x64 can play MWO usualy on Dx11 Very High (and i mean Very High in all Options) in 1920x1200 (and even better with the new 64 Bit Client where the FPS numbers are from).

In Public Matches i have ~40 to ~150 FPS.

In CW atm i have changed "Effects + Shadows" to medium and "Particles" to low and can now run CW with ~35 to ~80 FPS.

Everything else is still very high + PostAA.
Cocpit Glass, DoF and Filmgrain is disabled per User.cfg for all kind of matches.
Blurr + Damage Glow + VSync is off per ingame Options for all kind of matches.

If they optimize the HUD they may find another 10 FPS that would make me happy

PS: My old PhenomII x4 975 + HD6950 (32 Bit Client) could run Public Matches at 30 to 60 FPS at High settings.

Of course in some Forums People start whine bcs their Super-Rig cant run games at 500 FPS they need to have a smooth experience...no further comment!

Edited by Thorqemada, 25 December 2014 - 06:51 AM.

[Catamount]

40 to 150... except when it's 20

I, frankly, would be happy with minimums of 60 (or 50, or hell, 40), but if you ever find a way to get 500, do share. My epeen will thank you.

[Thorqemada]

The only 20s i have is when the Heat Vision Bug occurs or the 32 Bit Client suffers a Memory Leak - both have yet to happen in the 64 Bit Client.

PS: I also had so far no match were a specific Mech was a FPS breaker (the FPS arise when these specific Mechs get killed).

PPS: At every start of a match after Betty stopped Bitching you have so sweep your view left and right over where the enemy starts so it loads all Mechs into RAM or whatever - but you dont Benchmark a Level until it has fully loaded.

Edited by Thorqemada, 25 December 2014 - 11:59 AM.

[Catamount]

I think you'd find FRAPS recording 20s a great deal more than that, but it's hard to say what, exactly, is the biggest culprit. I've found it's largely map dependent, but the reason isn't apparent. Older maps generally perform worse than newer maps I've found. In fact, I've had Forest Colony runs with recorded minimums in the mid 20s on my 4.4GHZ 3770K (26 and 29 on the two runs I have records on hand of). That's vastly worse than anything I see in CW even.

Edited by Catamount, 25 December 2014 - 02:51 PM.

[Goose]

Forest hast lots of water; Frozen has them snow storms (read: particles); I forget if River is all that bad, anymore; But the rest of them are rather easy …

Edited by Goose, 25 December 2014 - 04:11 PM.

[Catamount]

Derp. You're right about Forest, and I never thought of that.

Of course, I seem to recall having no problem on Crimson Strait (don't recall River performance), so I think it's a combination of both literal level attributes and a situation where newer maps maybe have some better optimization thrown in somewhere.

[Goose]

I thought for months it was the trees swaying on the physics thread …

[Thorqemada]

Did some Fraps Benchmarks while leveling my CN9-AH (what a chore to play a Basic Mech and Boy the Cent has become such a Weaksauce - i get looked at and my CT is gone...)

Win7 x64
Gigabayte GA-990XA-UD3 rev.1.x Bios F14e
AMD FX8350 (stock)
Sapphire R9 280X Vapor-X OC (stock - Catalyst Omega 14.12 Standard Settings/Application Controlled)
2x 4GB Kingston Blu DDR3 1600
X-FI Titanium (Game Mode)
2x 1TB Samsung Spinpoint F3R Hard Disks - no SSD - MWO is installed on the second HD

MWO 64Bit
1920x1200 Dx11 Fullscreen
Blur, VSync, Damage Glow = Off
Cockpit Glass, Filmgrain, DoF = disabled in the user.cfg
Effects = Medium
Object Detail = Very High
Particles = Low
Post Processing = Very High
Shading = Very High
Shadows = Medium
Texturing = Very High
Environment = Very High
PostAA


Canyon:
Frames: 13142 - Time: 240000ms - Avg: 54.758 - Min: 31 - Max: 88

Caustic:
Frames: 13489 - Time: 240000ms - Avg: 56.204 - Min: 35 - Max: 107

Forrest Colony:
Frames: 15414 - Time: 240000ms - Avg: 64.225 - Min: 36 - Max: 102

HPG Manifold:
Frames: 14629 - Time: 240000ms - Avg: 60.954 - Min: 30 - Max: 108

River City:
Frames: 18543 - Time: 300000ms - Avg: 61.810 - Min: 34 - Max: 108

Viridian Bog:
Frames: 17745 - Time: 300000ms - Avg: 59.150 - Min: 34 - Max: 125

Initialy i wanted a 300 second Benchmark but the matches most of the time be decided at the 4 Minute after the first shot mark and so i went down to 240 second.
I started the Benchamrk after ~40 seconds in the game.
Had 1 weird bug where it showed me to have a Conquest on Caustic but then i started in River City.
Most of the time i played in normal view - did not test if Heat or Night Vision have a significant Impact on the FPS.
I am pretty happy with the FPS - Mining Colony did not show up - that is the Map where i expect some of the highest Numbers.
I did not run other Software like Sykpe or TS - it was pure Pugging Solo Drops.
I did run a Youtube Music Video (Santana - Tokyo Concert 2000) in the Background when i did the 240 second Tests.

Edited by Thorqemada, 25 December 2014 - 09:19 PM.

[Catamount]

Some of those settings are clearly making a huge performance difference. Why damage glow? I see it often mentioned. Does it really make that big a difference?

Also, just how do you remove film grain? DERP nevermind

Edited by Catamount, 25 December 2014 - 10:29 PM.

[Thorqemada]

Damage Glow - i switched it off and never missed it, i do actualy not know if it does make a performance difference.

Some stuff that works in the Testing Ground has no significant Impact in CW and Public Matches - like Object Detail that makes a big dfference in the Testing Ground but almost no difference in CW and PM.

Edited by Thorqemada, 25 December 2014 - 11:28 PM.

[Goose]

Catamount, on 25 December 2014 - 10:13 PM, said:

Why damage glow? I see it often mentioned. Does it really make that big a difference?

Somebody Important Said So

[Catamount]

Works for me!

[Thorqemada]

And a CW attacking on Boreal against IS Forces from the same match first and third wave - Settings as in my other post.

2014-12-27 03:57:15 - MWOClient
Frames: 13923 - Time: 240000ms - Avg: 58.013 - Min: 10 - Max: 139

2014-12-27 04:14:13 - MWOClient
Frames: 10532 - Time: 240000ms - Avg: 43.883 - Min: 23 - Max: 99

[Lord Letto]

Thorqemada, on 26 December 2014 - 07:22 PM, said:

And a CW attacking on Boreal against IS Forces from the same match first and third wave - Settings as in my other post.

2014-12-27 03:57:15 - MWOClient
Frames: 13923 - Time: 240000ms - Avg: 58.013 - Min: 10 - Max: 139

2014-12-27 04:14:13 - MWOClient
Frames: 10532 - Time: 240000ms - Avg: 43.883 - Min: 23 - Max: 99

Holly **** those Mins, OC that 8350 to Oblivion or Switch to Intel

[Catamount]

I'm almost thinking the 10 was a fluke or something. 20 I can believe for an 8350, but 10?

[Smokeyjedi]

LOL having fun tonight with my octocore....in the AM i should post some fun stuff......7 cores pegged @ >80% utilization whilst the remaining core did nothing untill particles and water were involved like my dedicated psyX GTX460.......does for BL2.....