Unigine Heaven 4.0 Benchmark Scores

[Win32]

Hi all!

I just got my new gtx 980 phantom today and i never used benchmark so this is the first time.
I just wonder if u folks can tell me that this results is ok or not.

 

[Ja.KooLit]

night.fox writes:
> So I figured, its my cpu is the culprit. Could be be also pci lanes limited. ...

Indeed, hard to tease out these issues, though normally one would expect a 4770K @ 4.4 to be more than sufficient.


> I have itching for x99 now. But ddr4 prices is slowly coming down now and I am still not decided which x99 mobo I will buy

I didn't bother with X99 due to the insane cost of DDR4, and the minimal gain one would have over X79. I bought
a bunch of X79 items at extremely good prices, total cost massively less than an X99 build:

ASUS R4E: 103 UKP (eBay, manufac. refurb BIN, item 281555191839)
Backplate: 10 UKP (eBay, new BIN, item 251788805619)
3930K C2: 225 UKP (eBay, used, item 261733380939)
1kW PSU: 55 UKP (ebay, used, item 231411474846)
Antec 302: 45 UKP (eBay, new, item 261735995836)
H80: 45 UKP (eBay, manufac. refurb BIN, I bought five of them in Apr/2013 from Scan)
Samsung 850 Pro: 107 UKP (eBay, new, item 221608679847)

The only items I bought new from normal sources were the RAM (GSkill TridentX 4x4GB @ 2400, 129 UKP from ebuyer),
fans (7x NDS PWM, total 75 UKP; I never use stock case fans, they suck), DVDRW (13 UKP), and the EVGA GTX 980 ACX 2.0 (455 UKP
from Scan). Late edition last night, bought a 3-pack of 4pin 30cm fan extension cables for 7.30 UKP (eBay item 171643026948),
so I can link up the side panel fans in a manner which means one doesn't have to fiddle with the mbd connections directly when
removing them for maintenance, etc.

(I try to get all sorts of items from eBay & elsewhere that, if possible, are new but sold via normal auction, etc. This week I bagged
a new OCZ ARC 100 240GB for only 43)

What would it cost to build a 6-core X99 with 16GB, new PSU, new 850 Pro, etc.?

This is my biggest gripe with X99, it's not enough better than X79 to be worth bothering with IMO unless cash really doesn't
matter and/or the 8-core will definitely help one's task (certainly wouldn't for gaming). I'd rather save the pennies, use them
for something else, which I did: CH Pro Throttle, Thrustmaster T-16000M joystick, and a Corsair H2100 headset, and will
probably buy a TrackIR later. Anyway, the savings by building X79 meant I could afford the peripherals that'll make the
gaming experience all the better. 8)

Ian.

PS. I typed this post a couple of hours ago, had to wait for the site to come back up before submitting. In the meantime,
I got sucked in testing Crysis on the new setup, max detail at 1080 gives around 65 to 160fps (4x AA, AOC, all in-game settings
maxed). Ahh the joys at last of Vsync ON.

Yeah if I am in US buying in ebay wouldnt be a problem since I could easily return. if there is a problem. and selling again should be no problem. The thing was 16gb G.skill 2400mhz DDR4 here cost about 350$. 16gb ddr3 2400mhz is 240$ so yeah 100$ difference.

x79 and x99 mobo and cpu almost same cost (yeah here even old tech price never go down) so between the 2, I would go for x99 since its newer tech. Ram only now. but the "naked" (without the heatsink), namely crucial, teamgroup, samsung etc. ddr4 8gb 2133mhz cost only 10$ higher than 8gb ddr3 1600mhz. So price of the ddr4 now is stabilizing. I could also see amazon ddr4 prices starting to drop since alot of manufacturers are opting for ddr4.

Of course if you have x79 already there is no point of upgrading to x99. But since I am going for HEDT platform, I would opt for newer tech.

ps, yeah seems TPU was down earlier

 

[mapesdhs]

Alas the markups over here can be nasty, and note I personally wouldn't buy the 5820K, I object to the
PCIe compromise (because IMO it nulls Intel's decision making process in releasing the 4820K, a chip
which I rather like and, obviously, is perfectly capable of outclassing the 5820K for gaming since it can
by definition handle 3/4-way CF/SLIno problem).

The Rampage V Extreme is 345 UKP here, so I saved 235 UKP (over $400) on the mbd alone. That one
saving covered the cost of almost all the peripheral extras. Likewise, the 5930K is 450 UKP, so that's
another 225 UKP saving, and the RAM was about 80 less than buying the DDR4 kit I would have chosen
if I had done an X99 build (16GB/2400 CL14).

I'll be more impressed with DDR4 when pricing comes down to the point where a 32GB kit is reasonable,
as it was for DDR3 a year ago.

Note that I know there are cheaper X99 boards, but since I bought an R4E I believe the comparison is fair.
Plus, there are equivalently cheaper X79 boards too, eg. I bought an ASUS P9X79 Deluxe for my brother
for only 75, fully boxed with all the original parts, works perfectly.

I do want to get an X99 setup eventually so I can expand my benchmarking stuff, but I'll wait until either
prices drop a lot more, or I can bag an eBay bargain (it happens, eg. I bought a 3970X for only 245 last Sept.)

Ian.

 

[Chatelike]

Hi all!

I just got my new gtx 980 phantom today and i never used benchmark so this is the first time.
I just wonder if u folks can tell me that this results is ok or not.

View attachment 62440

If you turn vsync off you will see actual results, around 1800

 

[Win32]

Today i will get new board,pc and ram and maybe then i will get full potencial because this is weird.max temp 55C...
If this temp is trully ok, then im very very satisfied! On idle 20C lol . The previous card was gtx 470 good edition and the temps was on max 70 with afterburner fan control and without, 90!

 

[Win32]

Anyway, new PC components epic fail...

 

[FX-GMC]

Anyway, new PC components epic fail...

View attachment 62477

As the post above referenced, make sure VSYNC is off. Your Max FPS should be above 140, not 61.5.

Also, Updated.

 

[Win32]

Thanks for reply but im trying to get some help here http://www.techpowerup.com/forums/threads/gtx-980-970-club.205567/page-13

It seems that i have unstable motherboard or fx-9370 cpu.

 

[Ebo]

thats my score on a I7-5820K, all at stock

 

[Ja.KooLit]

@FX-GMC it looks like your table for the multi-gpu score has an error. From what I recall @mapesdhs has 4700+ score and not more than 5700+ if he is indeed above 5700, should he be on top?

 

[Ebo]

thats the same settings as yours win32, even with screen set at the same.
All at stockspeed and on air.

 

[mapesdhs]

@FX-GMC it looks like your table for the multi-gpu score has an error. From what I recall @mapesdhs has 4700+ score and not more than 5700+ ...

Well spotted! Yes, it should be 4789, I guess though still @ no. 2 spot.

Will test shortly with a 2700K/5.0, should be interesting...

Ian.

 

[Ja.KooLit]

Well spotted! Yes, it should be 4789, I guess though still @ no. 2 spot.

Will test shortly with a 2700K/5.0, should be interesting...

Ian.

well when I check after he said he updated, i saw the list. I know and I cannot forget that 2nd slot because it was my spot for few months until you post your own results

so yeah I couldnt miss that part.

 

[krusha03]

Hmmm.... why is my score so pathetic? According to the table in the first post even a 7850 scores twice as much. This is a with the 7950 @ 900/1250 (stock clock) and +0% power limit. GPU-Z shows 100% gpu load at all times and no underclocking. Temp is at around 64C

 

[Ja.KooLit]

Hmmm.... why is my score so pathetic? According to the table in the first post even a 7850 scores twice as much. This is a with the 7950 @ 900/1250 (stock clock) and +0% power limit. GPU-Z shows 100% gpu load at all times and no underclocking. Temp is at around 64C

View attachment 62500

indeed thats a pathetic score. Have you tried updating drivers? mobo bios? and so on? what about windows update? have you tried re running and see if still same?

 

[krusha03]

indeed thats a pathetic score. Have you tried updating drivers? mobo bios? and so on? what about windows update? have you tried re running and see if still same?

Ok found the problem. I was toying with the application profiles to reduce the tearing in one older game but apparently i changed the global settings instead of only that game. Getting about 830 now which seems allright. since the only demanding game i play is BF4 with Mantle i havent noticed this affecting me otherwise

 

[dcf-joe]

I'm back to claim the number four slot. I have adjusted a lot of values in my BIOS other than core and memory clock speeds. Here is my new score with a i7 2600K at 4.8 GHz and my GTX 980 (Gigabyte G1 Gaming) at 1586 MHz core and 1966 MHz memory.

 

[mapesdhs]

That's not something I've tried at all yet, upping the memory clock. How did you find that to be in terms of
improved scores for a given increase? With previous cards, raising the core clock gave the best improvements.
Does the 980 behave the same way?

Ian.

 

[dcf-joe]

In my honest opinion, increasing the core still yields the better gains versus memory clock for clock. I adjust my memory when I want to get an extra few points.

 

[MiataManiac]

Hi all, just spent a couple hours combing through all this and it made me want to re-bench my system. Running an i5-2500K at stock speeds (3.3 Ghz) and two EVGA GTX 660 Ti SC+ cards in SLI (1,010/1,089). I'll have a new benchmark for a new system in a week or so. Figured this would be a great way to see before and after results. On another note, I'm the top 660 Ti SLi on the board... (I'm the only 660 Ti SLi on the board... I think. I'm kinda really tired so I could be completely wrong).

 

[xccel]

hello my setup R9 290 x3 not 3x R9 290x ^^

 

[mapesdhs]

Exploring PCIe issues, here are my three 980s, again at 1366MHz (the same minor +100 bump over stock), on an ASUS Maxiumus IV Extreme,
with a 5GHz 2700K (16GB RAM @ 2133 CL9):

It is less than the R4E setup, but not by much.


And just to provide a baseline data point on the single-GPU table, here's one 980 @ 1366MHz on the same M4E rig:

Hoping to test with an ASUS P7P55 WS Supercomputer and an i7 875K next, but after that the 980s go their separate ways,
two into gaming systems. I wanna play E.D. asap!

Ian.

PS. Mod-prod: main multi-GPU table still has the wrong number for my 3930K result.

 

[Ja.KooLit]

Exploring PCIe issues, here are my three 980s, again at 1366MHz (the same minor +100 bump over stock), on an ASUS Maxiumus IV Extreme,
with a 5GHz 2700K (16GB RAM @ 2133 CL9):

It is less than the R4E setup, but not by much.
Hoping to test with an ASUS P7P55 WS Supercomputer and an i7 875K next, but after that the 980s go their separate ways,
two into gaming systems. I wanna play E.D. asap!

Ian.

PS. Mod-prod: main multi-GPU table still has the wrong number for my 3930K result.

Oh boy, you bring me down one more slot . Curse those 980's lol

980's are really a beast.

But it wont make me still want to replace my 290's. I like them as now although I have more itch now for extreme platform

 

[BarbaricSoul]

@FX-GMC , seems to be a mistake on the single-gpu table, that number 4 spot score was done with a GTX 980, not a 970.

I used this debacle with the 970 misinformation to facilitate a "Step-Up", if you will, to a 980
EVGA GTX 980 SC ACX2.0
1560/1960

4790k 4.7

Im eyeing 2K but I spent a solid 2 hours and I don't think it'll happen without more voltage. *edit -via hardmod

 

[FX-GMC]

Charts should be fixed. Thanks for letting me know.