• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

**Official 3DMark 'Fire Strike' Leaderboard**

Final run for this rig, managed to get 4.7 stable in FS. (Well untill another driver update happens and they bring out another card :))

i7 [email protected] Ghz Water Cooled
290(P) Flashed to X @1,222 Mhz/1,600 Mhz Watercooled.

AMD 14.9.2 drivers

SCORE
11845 with AMD Radeon R9 290X(1x) and Intel Core i7-3770K Processor
Graphics Score 14137
Physics Score 12191
Combined Score 5245

http://www.3dmark.com/fs/3187129
 
Last edited:
Kaapstad said:
Lets look at this again

My quad 980 result was just a quick run and I used the first result I got.

If we are comparing graphics scores remember my CPU was only clocked @4.5 which is very low for 4 GPUs compared to 4.95 for Matts CPU.

If I shifted my 980s over to my 3930k PC which can do 5.1 (even more if I hook up my chiller) and maxed the cards I could probably get another 4 to 5k on the graphics score lol.

1080p is not the strong point for quad 290Xs, they need to be used at very high resolution to get that bus working.
Click to expand...

Since when did CPU clocks affect GS that much? And even if it did, shouldnt you have higher because of your 8 core? FS is fully threaded lol
 
Kaapstad said:
With a single card it does not effect the result too much but running a low CPU clock with 4 cards makes a very big difference.

Using 4, 6 or 8 cores with or without H/T makes very little difference to the graphics score result and clock for clock there is very little difference if you are using Haswell, SB, IB, Haswell E, IB-E or SB-E. The graphics score is all about raw CPU mhz to drive the cards.

The only place the extra CPU cores make a difference is on the physics and combined tests.
Click to expand...

Yea, but Haswell-E has 7-9% IPC advantage over SB-E. And your CPU is clocked to 4.5? 4.5 x 1.09 = 4.905GHz in terms of sandy. Which is about what Matts CPU is clocked to, not to mention you got DDR4 and 2 extra cores.

What am i missing?
 
Here's an hypthesis...

Kaapstad said:
If we are comparing graphics scores remember my CPU was only clocked @4.5 which is very low for 4 GPUs compared to 4.95 for Matts CPU.

If I shifted my 980s over to my 3930k PC which can do 5.1 (even more if I hook up my chiller) and maxed the cards I could probably get another 4 to 5k on the graphics score lol.
Click to expand...

So basically what you're saying is that with your 3930k at 5.1 - 5.2 even would give you another 4-5K GS. And that equals to about 4.6-4.7GHz Haswell-E.

Does that mean... If i can push maybe 50-100mhz more on my 3930k i can get 750-1250 more GS (your claim of acquired GS divided by 4~ because i have 1 card and you have 4)?

Hell, even 500 GS more would be great.

Edit: Or does this only apply to 4 cards? If so care to elaborate? :)
 
Last edited:
Kaapstad said:
It only makes a difference with multi GPU setup the more GPU grunt you have the higher you need to run the CPU to avoid a bottleneck simple.

As to using a single card if all you are after is maximum graphics score then lowering your CPU clockspeed can help as it will run slightly more efficiently and give you a few more points.
Click to expand...

Alright. Mind responding to this?

Kaapstad said:
With a single card it does not effect the result too much but running a low CPU clock with 4 cards makes a very big difference.

Using 4, 6 or 8 cores with or without H/T makes very little difference to the graphics score result and clock for clock there is very little difference if you are using Haswell, SB, IB, Haswell E, IB-E or SB-E. The graphics score is all about raw CPU mhz to drive the cards.

The only place the extra CPU cores make a difference is on the physics and combined tests.

Rangerjr1 said:
Yea, but Haswell-E has 7-9% IPC advantage over SB-E. And your CPU is clocked to 4.5? 4.5 x 1.09 = 4.905GHz in terms of sandy. Which is about what Matts CPU is clocked to, not to mention you got DDR4 and 2 extra cores.

What am i missing?
Click to expand...
Click to expand...
 
Edit: Oh i didnt see your post. Deleted.

Kaapstad said:
The extra efficiency makes hardly any difference on the synthetic benches 4.5 on Haswell E is near enough the same as 4.5 on SB E

It is all about raw mhz with 4 cards
Click to expand...



So... FX9590 5GHz or above Raw MHz would actually be better here by your logic?
 
Last edited:
Kaapstad said:
I don't know how it works with AMD CPUs.

With Intel CPUs on the graphics tests it is down to having enough speed to avoid bottlenecking the cards on the graphics tests (in these benches the graphics tests are solely concerned with the GPUs).
Click to expand...

But speed is DIRECTLY related to efficiency on the CPUs (IPC). And when the 4.5 haswell matches a 4.9 sandy because of the IPC difference you cant use the "i have lower clocks" excuse for your low GS.

Because as you see in this score: http://www.3dmark.com/fs/2893001 The GS is ridiculously close again, a little lower than matts because of the lower RAM clock but its the same exact CPU that you are using at 4.5GHz and i cant see any 4-5k GS difference compared to matts 3930k score :confused:.
 
Last edited:
Kaapstad said:
Sorry but I have tested what I posted above many many times on a lot of different CPUs in different configurations with x number of cores enabled and with and without H/T.

Also speed is not related to CPU efficiency.

This is the last I will post on this topic as a bench thread is not the right place for this sort of thing.
Click to expand...

Heh, for all this time you could just admit that even you can see the scaling is We have a swear filter for a reason, but for some reason you kept digging a hole.

And instead of proving what seems like a ridiculous claim and i have never even heard anyone else say before, you just choose to stop talking about the topic.

Credibility:

32b0504811.jpg

I havent really seen anyone go as far as claiming to be at a disadvantage with Haswell-E 8 core at 4.5GHz in BENCHMARKS.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom