Hello everyone,

 

I am on the fence of weather to upgrade. This looks like what I am leaning towards.

 

Intel Core i7-3930K Sandy Bridge

 

Thanks Techguy, Dazz, and Word Not Allowed for all the testing and help you have provided...!

If you want great OC'ing get the 2700K.

 

But, if you are willing to get less OC'ng but really good memory performance get IB.

 

all right I am going to ask the obvious. For FSX, which is better. the marginal difference in OC number or the memory performance difference?

 

:Loser:

all right I am going to ask the obvious. For FSX, which is better. the marginal difference in OC number or the memory performance difference?

 

:Loser:

 

6 dozen and half of the other...only concern I have is to get IB to go 4.7ghz, it seems is rather rare as well as can be too hot, and not eveyone will be able to get fast ram in excess of 2133, such as 2400 stable and sourced... But I think 4.8-5 ghz for SB, as others have said, will be like 4-6-4.7 on IB...so I think both systems will be very close in FSX as seen from Techmax results comparing to the FSXmark averages. PCI 3.0, I don't think will be a cure for microstutters etc etc as I get them once in awhile and only fly with autogen off with photoscenery only, so I know the bottleneck is still cpu related not the PCI bus as my bus is nowhere near saturated...

deleted

Sorry for the lack of updates this weekend, been very busy and expect to be busy for the next couple days at least as my wife and I will be celebrating our 1 month "anniversary" tonight and then tomorrow we will be putting a new desk in the office to accommodate all the new PC equipment. If we get done sooner than expected I shall resume testing. Otherwise, I probably won't have time again until Thursday (Wednesday night = Bible study). I'll let you guys know as soon as I have more info. I still plan on testing PCIe 3 performance, though I won't be able to actually knock it down to PCIe 2, I can simulate it with PCIe 3 x8 speed by putting in another card as dazz suggested. I'll just disable the card in Windows so the board will still be forced to run the slots at half speed but I won't have to worry about which card is selected by FSX.

Next month probably after Computex......DDR3 2800 MHz

 

 

 

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

4.9GHz HT Off 1.41v

 

I guess IB doesn't run @ low volts!

If you want great OC'ing get the 2700K.

 

But, if you are willing to get less OC'ng but really good memory performance get IB.

 

Manny et al,

 

Let's just say you can for sure do 5Ghz w/ SB using air or water, and 4.7Ghz w/ IB using air or water. You are looking at a whopping 6% increase in clock speed. In the most demanding situations in FSX on these platforms where your system is very stressed to keep up w/ the workload, you are seeing for example 22FPS. The difference is going to be a grand total of 1.3FPS. In no way would you be able to discern this difference were it not for a frame counter. So, the only meaningful argument is in smoothness or texture update rate and hopefully memory speed or PCIe 3 will impact this area. Hopefully something measurable will demonstrate SB v IB in this regard, to help folks decide if the sort of temps involved are worth it, as certainly clock speed isn't going to be meaningful . . . beyond braggin' rights that is to say ;o)

 

Edit: sorry Manny, I see you asked this question later in the thread. One would love to know if intel can improve on the heat issue w/ Haswell. I can't imagine them doing anything w/ IB as far as changing to fluxless solder since anyone w/ earlier releases who cared would demand a replacement! Interesting they now have their over clockers' warranty or whatever it was.

 

Next month probably after Computex......DDR3 2800 MHz

 

 

 

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

 

Temps?

IB as far as changing to fluxless solder

 

Someone already de-lidded an IB chip and it made almost no difference on the temperatures. The water block was directly in contact with the CPU die.

Someone already de-lidded an IB chip and it made almost no difference on the temperatures. The water block was directly in contact with the CPU die.

 

I believe it Ben--I couldn't see why they would do this AND offer a replacement warranty for over clockers if it mattered at all.

Someone already de-lidded an IB chip and it made almost no difference on the temperatures. The water block was directly in contact with the CPU die.

 

PC Watch out of Japan opened one up and replaced the timm with better timm. That lowered the max OC temp by 15C to 20C depending on the timm used.

 

PC Watch article http://pc.watch.impress.co.jp/docs/column/sebuncha/20120511_532119.html

English summary of the original article http://www.techpowerup.com/165882/TIM-is-Behind-Ivy-Bridge-Temperatures-After-All.html

There are three or four English summarys from different sources out in the wild, you can search for them if you want to read the others.

 

Here is a link to someone at xtremesystems.org that did his own timm replacement test, his temps dropped 8C and his ambient temp with the 8C drop was 3C higher than when he tested the stock timm.

http://www.xtremesystems.org/forums/showthread.php?280892-3770k-IHS-Removals-CPU-temp-dropped-from-79C-to-71C

PC Watch out of Japan opened one up and replaced the timm with better timm. That lowered the max OC temp by 15C to 20C depending on the timm used.

 

PC Watch article http://pc.watch.impr...511_532119.html

English summary of the original article http://www.techpower...-After-All.html

There are three or four English summarys from different sources out in the wild, you can search for them if you want to read the others.

 

Here is a link to someone at xtremesystems.org that did his own timm replacement test, his temps dropped 8C and his ambient temp with the 8C drop was 3C higher than when he tested the stock timm.

http://www.xtremesys...from-79C-to-71C

 

There is alot of people who have tried to remove the stuff and had no effect on temps, but there is always some people who do things that others can't; I would not bother...

There is alot of people who have tried to remove the stuff and had no effect on temps, but there is always some people who do things that others can't; I would not bother...

 

Not so far as I've seen. Thus far it was just one user and he didn't appear to use a proper mounting method to ensure proper contact. All the other reports have indicated significantly decreased temps.

New numbers w/out HT (lower voltages & temps) but with another PCIe card installed to force the 680 to run at PCIe 3.0 x8 speed (simulating 2.0 x16 speed):

 

 

Frames, Time (ms), Min, Max, Avg

14006, 300000, 28.33, 65, 46.36

 

No real difference measured at this time, though if I comment subjectively I would have to say that the full speed PCIe runs seem smoother. I perceived occasional hitching (micro-stuttering) running @ x8 that was not there @ x16. Also, I need to go back and re-run the x16 numbers now with HT off and lower voltages to see if the lower temps allow the system to maintain full turbo speed for the entire length of the benchmark. I don't know if I'll have time for a full report on this tonight, but I'll try.

Got test results back at x16 speed:

 

 

Frames, Time (ms), Min, Max, Avg

14080, 300000, 28, 67, 47

 

So nothing big, but a minor bump and an observable reduction of micro-stuttering (even to the point of elimination IMHO). I would say that texture and autogen loading are also faster but that's harder to quantify than fluid motion. Only thing left to do at this point is OC the graphics card, which I don't expect to produce much of an increase in FPS. At this point I'd have to say I'm pretty happy with the results. It's not the absolute fastest system on the FSXMark 11 list, but it's darn close and at a good 500MHz slower, so that says something.

 

I hope this has been and will continue to be useful information to anyone considering an upgrade/new system build in the next several months. I'll post final results once I OC the GPU here sometime later this week.