LTT 3DMark Thread

[VagabondWraith]

4 minutes ago, done12many2 said:

 

You need to push that 6700k further too.  

For sure. I haven't paid much attention to CPU overclocking (even though I have a Z board and K CPU), but I may look more into it as I have been wanting to do so. I'm quite busy and CPU overclocking takes much more time to be stable. Why can't it be like GPU overclocking? Lol?

Just a few clicks of Afterburner and you're good to go.

 

Would you recommend a quick OC profile from Gigabyte for a quick run? Like say 4.5-4.6 area. I would really have to sit down one weekend and do self overclocking when I get the time.

[done12many2]

1 minute ago, VagabondWraith said:

Would you recommend a quick OC profile from Gigabyte for a quick run? Like say 4.5-4.6 area. I would really have to sit down one weekend and do self overclocking when I get the time.

 

Yeah, sure.  Just check and double check your voltages with the auto overclocks as they tend to be high.  You can just use it as a starting point and tweak things from there.

[VagabondWraith]

4 minutes ago, done12many2 said:

 

Yeah, sure.  Just check and double check your voltages with the auto overclocks as they tend to be high.  You can just use it as a starting point and tweak things from there.

Thanks buddy. I'll try that. As long as my volts are around 1.3ish I'll be fine. I'll definitely monitor voltage as I don't want a spike to like 1.8 or something. Haha

[done12many2]

1 minute ago, VagabondWraith said:

Thanks buddy. I'll try that. As long as my volts are around 1.3ish I'll be fine. I'll definitely monitor voltage as I don't want a spike to like 1.8 or something. Haha

You're welcome bud.

 

You can go higher than 1.3 for benchmarking.  Just watch temps and go as high as you can.  It's only for a short time to get the higher physics score.  Then you can set it lower for daily use.

[done12many2]

12 hours ago, done12many2 said:

Benchmark:  Fire Strike
CPU:  i7 7700k @ 5.4 GHz
GPU:  2 x GTX 1080
GPU Core:  2,100 MHz
GPU Memory:  5,805 MHz (11,610 MHz effective)

Score:  29,496
3DMark Link:  http://www.3dmark.com/fs/11782039

 

No system memory tweaks other than selecting a higher operating frequency while using timings from 3200 XMP.  More to come on that and GPU clock speed. CPU is a simple 54x and go.

 

 

Update:

 

Benchmark:  Fire Strike
CPU:  i7 7700k @ 5.4 GHz
GPU:  2 x GTX 1080
GPU Core:  2,100 MHz
GPU Memory:  5,940 MHz (11,880 MHz effective)

Score: 29,879
3DMark Link:  http://www.3dmark.com/fs/11787050

 

 

[DildorTheDecent]

6 minutes ago, done12many2 said:

 

 

 

wat

[done12many2]

2 minutes ago, DildorTheDecent said:

 

wat

 

I'm just trying to keep up with those bad ass Titan XPs.  

[Vellinious]

1 hour ago, done12many2 said:

 

I'm just trying to keep up with those bad ass Titan XPs.  

No kidding.  Have you see this?  Not sure what magical fairy dust this guy is sprinkling on his GPUs, but I wish he'd freakin share.  49k?  /smh

http://www.3dmark.com/compare/fs/10372425/fs/11379496

 

[Jumper118]

2 minutes ago, Vellinious said:

No kidding.  Have you see this?  Not sure what magical fairy dust this guy is sprinkling on his GPUs, but I wish he'd freakin share.  49k?  /smh

http://www.3dmark.com/compare/fs/10372425/fs/11379496

 

is his score valid? 

[Vellinious]

6 minutes ago, Jumper118 said:

is his score valid? 

Says it is.....I had to roll my rig outside, lower my coolant to 2c to get mine to hit 48k, and he's got his hitting 49......I've tried everything I could think of, and can't touch it.  (aside from flashing to the T4 bios or hard modding.....I won't do either)

http://www.3dmark.com/fs/10372425

 

[Jumper118]

4 minutes ago, Vellinious said:

Says it is.....I had to roll my rig outside, lower my coolant to 2c to get mine to hit 48k, and he's got his hitting 49......I've tried everything I could think of, and can't touch it.  (aside from flashing to the T4 bios or hard modding.....I won't do that)

http://www.3dmark.com/fs/10372425

 

i think his extra cpu speed nets higher gpu score at high fps. thats why. 

[Vellinious]

5 minutes ago, Jumper118 said:

i think his extra cpu speed nets higher gpu score at high fps. thats why. 

Ya know...I've never seen a difference in the graphics score in Firestrike no matter what my CPU is running.  I've heard people say that, but.....I just can't get there, as I've never seen it.  Some of my best graphics scores come during test runs, when my CPU is running my daily clocks.  /shrug

[Jumper118]

6 minutes ago, Vellinious said:

Ya know...I've never seen a difference in the graphics score in Firestrike no matter what my CPU is running.  I've heard people say that, but.....I just can't get there, as I've never seen it.  Some of my best graphics scores come during test runs, when my CPU is running my daily clocks.  /shrug

makes a difference for me running e5 e2670 v1 vs 2500K on a 7990 cf for gpu score. i suspect he has tweaked the os as well as i have seen his name on hwbot and i think he has a lot of points. 

[done12many2]

26 minutes ago, Vellinious said:

No kidding.  Have you see this?  Not sure what magical fairy dust this guy is sprinkling on his GPUs, but I wish he'd freakin share.  49k?  /smh

http://www.3dmark.com/compare/fs/10372425/fs/11379496

 

 

Yeah, there's a bit of trickery involved there.  I know folks are cheating in a lot more obvious ways as it's being talked about on OCN.

 

25 minutes ago, Vellinious said:

Says it is.....I had to roll my rig outside, lower my coolant to 2c to get mine to hit 48k, and he's got his hitting 49......I've tried everything I could think of, and can't touch it.  (aside from flashing to the T4 bios or hard modding.....I won't do either)

http://www.3dmark.com/fs/10372425

 

 

This is in my near future.  I'm either rolling it out into the garage or onto the rear deck.  If I put it on the rear deck, I can just plug it in outside and close the glass door.  Then use a wireless keyboard and mouse and I never have to get cold. lol

 

4 minutes ago, Jumper118 said:

makes a difference for me running e5 e2670 v1 vs 2500K on a 7990 cf for gpu score. i suspect he has tweaked the os as well as i have seen his name on hwbot and i think he has a lot of points. 

 

My best 5960x and 7700k graphics scores with the same SLI are almost identical, but slightly favoring my 7700k setup.  Obviously, the 5960x kills my 7700k in physics, but I do average 17.4k to 17.5k now with the 7700k, and that's not too shabby at all for 4 cores.

[Jumper118]

3 minutes ago, done12many2 said:

 

Yeah, there's a bit of trickery involved there.  I know folks are cheating in a lot more obvious ways as it's being talked about on OCN.

 

 

This is in my near future.  I'm either rolling it out into the garage or onto the rear deck.  If I put it on the rear deck, I can just plug it in outside and close the glass door.  Then use a wireless keyboard and mouse and I never have to get cold. lol

 

 

My best 5960x and 7700k graphics scores with the same SLI are almost identical, but slightly favoring my 7700k setup.  Obviously, the 5960x kills my 7700k in physics, but I do average 17.4k to 17.5k now with the 7700k, and that's not too shabby at all for 4 cores.

must be something to do with crossfire then because it makes a big difference on 7990/7970 and rx480 cf

[done12many2]

4 minutes ago, Jumper118 said:

must be something to do with crossfire then because it makes a big difference on 7990/7970 and rx480 cf

 

Does crossfire have more CPU overhead then SLI?  Maybe something related to that?

[Jumper118]

1 minute ago, done12many2 said:

 

Does crossfire have more CPU overhead then SLI?  Maybe something related to that?

yes. there is also some kind of issue with over 16 threads in the combined test where you get a better score with a 6950x with ht than on  

[Vellinious]

44 minutes ago, Jumper118 said:

yes. there is also some kind of issue with over 16 threads in the combined test where you get a better score with a 6950x with ht than on  

I have noticed some crazy ups and downs in the combined score.

 

1 hour ago, Jumper118 said:

makes a difference for me running e5 e2670 v1 vs 2500K on a 7990 cf for gpu score. i suspect he has tweaked the os as well as i have seen his name on hwbot and i think he has a lot of points. 

As long as I beat him, I guess it doesn't matter....irritates me though, that I had to ramp up the CPU to do it.

[Jumper118]

19 minutes ago, Vellinious said:

I have noticed some crazy ups and downs in the combined score.

 

As long as I beat him, I guess it doesn't matter....irritates me though, that I had to ramp up the CPU to do it.

If you used lod you could destroy any score.  Just wouldn't be valid

[VagabondWraith]

Any of you guys know the average for TXP overclocks? I was only able to hit 2076 on my 1080 so I'm pleasantly surprised to see 2126 on my TXP, even if it has 1,024 more cores. Would my OC be considered average or above average?

[done12many2]

7 minutes ago, VagabondWraith said:

Any of you guys know the average for TXP overclocks? I was only able to hit 2076 on my 1080 so I'm pleasantly surprised to see 2126 on my TXP, even if it has 1,024 more cores. Would my OC be considered average or above average?

 

Seems about right if it's cooled via hybrid.  To be honest, I wouldn't even worry about clockspeed much.  What matters to me is what it produces, if that makes sense.  One setup may run 2200 MHz which results in 200 FPS and another may run 2190 MHz and 201 FPS.  In that example, I'd take the 2190 MHz.   I only really care about the work produced.

[VagabondWraith]

7 minutes ago, done12many2 said:

 

Seems about right if it's cooled via hybrid.  To be honest, I wouldn't even worry about clockspeed much.  What matters to me is what it produces, if that makes sense.  One setup may run 2200 MHz which results in 200 FPS and another may run 2190 MHz and 201 FPS.  In that example, I'd take the 2190 MHz.   I only really care about the work produced.

So I added a +225 offset to the core to bring it to 2126. I added +240 but it wouldn't budge past 2126 even though it didn't crash. My score decreased at +240 so I assume 2126 +225 would be the max for my GPU?

[done12many2]

26 minutes ago, VagabondWraith said:

So I added a +225 offset to the core to bring it to 2126. I added +240 but it wouldn't budge past 2126 even though it didn't crash. My score decreased at +240 so I assume 2126 +225 would be the max for my GPU?

 

I would say so. Sounds like GPU Boost is doing what it does.  Taking things too far will often result in a slight drop once you've passed the point of what it determines is best. 

 

You also have the option of manually  setting your own voltage/clock speed curve, which can result in getting more out of the GPU benchmark wise.  For daily use, I find the the normal offset method is best and more flexible for different situations. 

[VagabondWraith]

1 minute ago, done12many2 said:

 

I would say so. Sounds like GPU Boost is doing what it does.  Taking things too far will often result in a slight drop once you've passed the point of what it determines is best. 

 

You also have the option of manual setting your own voltage/clock speed curve, which can result in getting more out of the GPU benchmark wise.  For daily use, I find the the normal offset method is best and more flexible for different situations. 

I'll check into that more later. Thank you.

[Vellinious]

13 hours ago, VagabondWraith said:

So I added a +225 offset to the core to bring it to 2126. I added +240 but it wouldn't budge past 2126 even though it didn't crash. My score decreased at +240 so I assume 2126 +225 would be the max for my GPU?

Cool it off more, and you'll likely be able to get a little more out of it.  Pascal hates heat....any kind of heat.