Ivy Bridge hits the streets--post your results here

[alainneedle1 97]

Tested Prime 95 this morning, temps hit up to 95 on the hottest core after 10-15 minutes of testing.

 

eeeeeeeeeeee .....any noticable smelling "kinda" burning rubber? :im Not Worthy:

[Alan_A 1,650]

Right now the hardware using PCI-e 3.0 are not mature enough, (remember when the first SB mobo came out...) wait for the GTX 685 +, driver for this card are to young also, same with the mobo, we need another 6 months...

 

Nick Needham over at Simforums says the same thing. I'm persuaded to wait until the last quarter of the year to let things sort themselves out. Glad to benefit from the heroism of pioneers but I'll hang loose until the territory is a bit more settled.

[HLJames 2,568]

Hi, I believe you had a SB system before for FSX, any comments, comparison with the new IB system that you current have?

 

Yeah my old P67 B3 board has finally weakened, loged too many hours, burned up too much jet fuel.

 

CPU:Core i7-3770k @ 4.9GHz HT Off 1.41v

Mobo: Asus Maximus V Gene

RAM: G.Skill Trident X 8GB 2666 MHz (11-13-13-32-1N)

GPU: EVGA GTX 680

HDD 1: WD Velociraptor X2 1.82TB intel RAID 0

HDD 2: WD Caviar Black X2 1.82TB intel RAID 0

PSU: Cooler Master SP Gold 1200 W

Case: Cooler Master Cosmos S

Cooling: Blackice Extreme 360 / Laing D5 / Swiftech Apogee HD / feser one

OS: Windows 7 x64

 

HLJAMES

[Anxu00 176]

Yeah my old P67 B3 board has finally weakened, loged too many hours, burned up too much jet fuel.

 

CPU:Core i7-3770k @ 4.9GHz HT Off 1.41v

Mobo: Asus Maximus V Gene

RAM: G.Skill Trident X 8GB 2666 MHz (11-13-13-32-1N)

GPU: EVGA GTX 680

HDD 1: WD Velociraptor X2 1.82TB intel RAID 0

HDD 2: WD Caviar Black X2 1.82TB intel RAID 0

PSU: Cooler Master SP Gold 1200 W

Case: Cooler Master Cosmos S

Cooling: Blackice Extreme 360 / Laing D5 / Swiftech Apogee HD / feser one

OS: Windows 7 x64

 

HLJAMES

So how is the FSX experience for you coming from Sandy to Ivy? I think everyone is dying to know that. TechGuyMax was going from pre-lynnfield area, so it's a difference experience.

 

Thanks

[ghetto 0]

So my new Ivybridge turned up this morning.. It was an oveclocked bundle from scan..

Mobo= asus P877-v Delux

I7 3770k clocked to 4.6ghz

8gb corsair vengance..

OCZ 240gb Vortex4..

 

Since I have now acumulated 3 corsair 120gb ssd drives (120gb was eaten up so quickly) I decided to try raid 0 on 2x corsair 120gb windows went in a treat but then I got to thinking why not 3x 120gb in raid 0 since they are all the same, so I did that..I dont think it worked so well as installing adobe reader and the asus drivers seemed to take forever..I went ahead and installed FSX and noticed straight away that the first splash screen was taking 30-40 seconds which has never been the case normally 10 seconds on my previous system.. I ran the fsxmark11 test anyway to see how it faired and I get average 28 FPS..

I am now reinstalling windows on the vertex4 drive and will try again..

 

Should have a gtx690 with me tomorrow for some more testing :)

[Ben Cap 89]

gtx690

 

Why in the WORLD would you buy a 1000 dollar GPU?!

[alainneedle1 97]

Why in the WORLD would you buy a 1000 dollar GPU?!

 

Web surfing.........

 

Seriously, it's probably because he can, the other reason maybe because he's playing other games, we all understand that a GTX 690 will do nothing for FSX FPS wise (may also run worse) but if he is flying at high resolution the extra GB on the card will probably help.

[dazz 69]

Web surfing.........

 

[HLJames 2,568]

Ivy Bridge cannot tolerate continuous 100% core utilization. Do not even think of using those multithreaded intense floating point utilities.

 

FSX core activity rises and falls in a pattern that permits inactive cores to act as heat sinks for active cores and in a frequency that permits cooling between cycles.

 

Thermal Event – As you step up in core frequency and measure performance in FSX core temps. Will average about mid 60*C. Performance will increase step by step. At some point one click and temperature will jump to 80*C and there will be no performance increase.

 

The location of this Thermal Event is a function of your cooling; Corsair H100 runs FSX OK @ 4.7GHz, Classic Water runs FSX OK @ 4.9 GHz.

 

Memory – Your CPU has to wait for memory to provide data, its called wait state. The slower the memory the more the wait state. Sandy Bridge is capped at 2133 MHz. In the meantime memory manufacturers have moved on to 2200, 2400, 2600, 2666, 2800, 2933, 3000, 3200 MHz. Ivy Bridge will use these memory speeds to force the memory sub system to work harder with less wait state.

 

HLJAMES

[dazz 69]

meh, I'm in wait state myself... waiting for Haswell

[JohnFromCO 15]

Why in the WORLD would you buy a 1000 dollar GPU?!

 

Because it is cheaper than 680 SLI?

 

Plus you might actually find a 690 in stock.

[ghetto 0]

Because a 690 according to the benchmarks I found is kicking my 590's butt :)

I game at 3840 x 720 in 3d vision surround so I need all the power I can get..

I was tied between going 680sli or 690 and John is right 580's are still in short supply, although apparently sli 670's are matching 90-95% of the 680sli performance so if cost was a factor I could have gone £700 for 2x gtx670's but I opted for gtx 690 for the extra £190 it cost, the thing is a work of art :)

 

Results in on my first stage of testing.. at first I could not get passed an average of 28fps.. Then I moved the monitor to one of the other card dvi outputs and manually changed Physx to that slot on the GTX590 .. I then scored an average of 42fps at 4.5ghz..

So nothing exciting there.. still have autogen constantly popping up and occasional flashes but a very slight micro stutter I used to get with the sandybridge seems to have gone..

 

Gonna have a delay now as windows is installing 106 updates :0 LOL

[Noel 4,502]

meh, I'm in wait state myself... waiting for Haswell

 

Me2 Dazz, and look what I'm running! I don't hold much hope for Haswell either, but who knows maybe the architecture change will matter. In the mean time I hope maybe something (Laminar, P3D) starts using modern hardware more efficiently, in which case I will pick up a nice 6-core solution would that will be utilized by newer sim platforms. I can tell you I have every bit as much fun flying as anyone here, even though I have to be a little mindful where I fly. I just use softer aircraft (Super MD-80, etc) when I'm flying ORBX scenery, or flying into KJFK. In lighter scenery I can fly the NGX just great. In fact I often jump from PHTO to PHLI complete w/ all the splendor from using GSX, in the NGX, and have totally excellent and smooth flight, and look what I'm running! Fortunately I don't need an array of wide screens to be happy ;o)

[ghetto 0]

Just ran the benchmark test using a amd powercolour 7970..

Average 42.8

comparing that against my 590 it seems the Amd has the edge..

While watching the benchmarks run on the AMD I notice that the flashes I get with the nvidia have gone but it also looks like anti aliasing does not work as well as with the nvidia..

Hopefully tomorrow I will run the tests with the GTX690..

[SAAB340 28]

I will upload the .cvs files when I get a chance, probably not tonight unless it takes a lot less time to re-do the office than I think it will. As for the benefits of PCIe 3 vs. 2, as I said in my observation of the numerous FSXMark 11 runs at various configurations, I would say that subjectively PCIe 3 is faster (again, simulating PCIe 2 by running at half speed) as is noticed by the reduction of micro-stuttering. I thought perhaps it was just a placebo at first but after observing and comparing numerous runs I can definitely say there are less (perhaps even no) micro-stutters. So while the FPS gains are very small, the effect is far more noticeable in the sim.

 

Max, would you mind logging frametimes and FPS on the first run of FSmark11 (the one you discard in the official result) with x16 and x8 bus and upload it?. I would love to analyze the frametimes for stutters. The important bit is that FSX is closed down fully between each benchmark run. That way the GPU memory gets flushed and all data needs to be transported via the PCIe bus to the GPU again each run. I suspect the difference in PCIe bandwidth will be notable. Just to remember, that will not give an official FSmark11 result but it will show if the improved bandwidth makes a difference during normal usage of FSX.

 

Thanks for your testing