New MSI GTX 1060 6GB Getting Lousy Fire Strike Score

[Sasqui]

Brand new MSI GTX 1060 6GB arrived Friday from Newegg. All Tested on the same 4970k system, clean driver install(s). Even installed the driver twice for the 1060 and used drivercleaner (or whatever they call it now) before re-installing. This is on Win 7 64 bit

Score with Sapphire 290x = 10,364
Score with ASUS 1070 = 14,758
Score with MSI 1060 6gb = 7,071 (5 seprate runs)

Physics score on all is right around 11,900 (I assume that is the CPU)

From all I've read, I should be getting above 11,000. Do I have BUM video card?

See screen below the graph spike in CPUz was during the 3DMark FS run.

CPUz:

 

[jboydgolfer]

I can't speak for why you would be getting less than normal results ,but I can say that for the last month or two I have not gotten a single result from 3-D Mark that doesn't give the invalid error( not able to verify score).... so you're not the only one having problems with that admittedly a different problem but problem just the same

it recognizes NOTHING in my PC, not the CPU, the GPU, the driver, the time...

 

[Mussels]

Checked the unlikely stuff, like your watercooling pump being disconnected and causing CPU to throttle, etc etc?

Run benchmarks and confirm the rest of the system is behaving as expected, double check the simple things like Vsync being correctly disabled, and so on.

 

[Sasqui]

Checked the unlikely stuff, like your watercooling pump being disconnected and causing CPU to throttle, etc etc?

Run benchmarks and confirm the rest of the system is behaving as expected, double check the simple things like Vsync being correctly disabled, and so on.

No watercooling in this rig, and nothing has changed AT ALL between the 1060 and testing the other two cards (1070 and 290x) - just a driver-reinstall each time.

 

[jboydgolfer]

I wonder if the 1060 has some power throttling feature built into the bios or the driver that's aggressively holding it back..... can you run after burner and set it to non-reference option, or possibly remove the temp restrictions ? Or increase fan speed to way above normal to keep it 65c or below? As mussels said maybe the temperature is holding back or causing throttling

 

[P4-630]

 

[jboydgolfer]

@Sasqui

there is Definitely something wrong, my Reference gtx 970 scores a 10,000+ running that same run of Firestrike, ill garuntee the results even tho i keep getting these "invalid score " errors... my PC is as posted below

i7 4790
GTX 970 *not in SLI*
Asrock Extreme 4 Z97
Windows 10 64b
16 Gb's Dog shit Kingston HyperX blu from 2010

a 1060 should be around 20% better +/- a bit,than my old GPU.

 

[Mussels]

No watercooling in this rig, and nothing has changed AT ALL between the 1060 and testing the other two cards (1070 and 290x) - just a driver-reinstall each time.

i spent two weeks figuring out a lag issue only to realise i took the power out to my water pump whilst leaving the fans running, hence that example... i7 2600's perform pretty well even at 100C XD

how about Vsync in the nvcp?

 

[jboydgolfer]

can u try a different bench? maybe superposition?

 

[Sasqui]

View attachment 89472

Noticed that too..

i spent two weeks figuring out a lag issue only to realise i took the power out to my water pump whilst leaving the fans running, hence that example... i7 2600's perform pretty well even at 100C XD

how about Vsync in the nvcp?

Vsync is off

can u try a different bench? maybe superposition?

Might as well, it'll be later. Off to the beach

 

[jboydgolfer]

Off to the beach

enjoy

 

[Vya Domus]

This is on my system , there dose seem to be something wrong , graphics score is almost half of mine. Make sure you didn't dialed back the power limit or messed with something within NV control panel.

 

[jboydgolfer]

View attachment 89473

This is on my system , there dose seem to be something wrong , graphics score is almost half of mine. Make sure you didn't dialed back the power limit or messed with something within NV control panel.

Did you have G sink Enabled during that test? I'm just curious because I was under the impression the 1060 should be faster ,but your score seems lower than I would've figured it would've guessed.

 

[Vya Domus]

Did you have G sink Enabled during that test? I'm just curious because I was under the impression the 1060 should be faster ,but your score seems lower than I would've figured it would've guessed.

Nope , I do not even have a g-sync monitor , it always gives me that warning , no idea why.

I just looked some scores and it seems like 13-15k is the range for 1060 with some odd ones here and there. I have always suspected Firestrike to not be very reliable.

 

[jboydgolfer]

Nope , I do not even have a g-sync monitor , it always gives me that warning , no idea why.

Hmm.... it seems like 3d mark likes giving odd messages. i actually figured out what was giving me the error of "invalid score" it was my intel iGPU was set to enabled i think, i disabled it, and now it gives me the Valid score....but im still a bit surprised at the 1060's scores....

 

[Vya Domus]

Hmm.... it seems like 3d mark likes giving odd messages. i actually figured out what was giving me the error of "invalid score" it was my intel iGPU was set to enabled i think, i disabled it, and now it gives me the Valid score....but im still a bit surprised at the 1060's scores....

Firestrike scores seem all over the place for most people. Can't find anywhere exact numbers for any card. Apparently a GTX 980 which is supposed to be match scores about the same.

 

[jboydgolfer]

yeah, likely due to varying configs....

 

[Vya Domus]

yeah, likely due to varying configs....

It makes sense for the combined and physics scores , but the graphics score shouldn't vary that much.

 

[rtwjunkie]

It makes sense for the combined and physics scores , but the graphics score shouldn't vary that much.

Same system with switching between a lower end CPU and a higher end, faster one, it will actually make a pretty good difference in the overall score, but, it can be noticed a bit in just the graphics score too.

Granted, this is all benchmarking way beyond most people's gaming settings, and most wouldn't know the difference in everyday gaming.

 

[jboydgolfer]

It makes sense for the combined and physics scores , but the graphics score shouldn't vary that much.

In my experience with testing various configurations in both CPU and GPU combination you Unigine superposition was basically 100% GPU dependent ....and it made no difference that I would use either a Pentium dual core or an i7.

I wonder ,if you look under Nvidia control panel if you have your 1060 set to handle the physics ?it just seems odd that your physics score would be that low.

Goes for @Sasqui also, if cpu is handling physx it kills the score unfairly. I've noticed Nvidia control panel even if you change it to the GPU for me it'll set back to CPU sometimes... but now that I look your Op physX score isnt low.

I bet if either of you guys set the benchmark to hire settings like firestrike extreme your scores would increase relative to the test

 

[Sasqui]

enjoy

Folks, two different 290x's, a 1070 and the 1060 were tested in this exact rig numerous times with stock 4970k speeds, results were consistent between card and on multiple runs see the OP, the only one with the shit score is the 1060.

Here's Hampton beach in NH this afternoon. Not a bad view

 

[Toothless]

I don't dare go outside in the sun for more than 30 minutes because my skin is neon white and burns way too easily.

On topic, I'm going to test my card and see if it matches other 1080TI scores because maybe it's an NVIDIA bug.

 

[Laki89]

I have dumb question for OP:

Did you plugged 6 pin power cable from PSU to graphics card before turning on pc?

 

[Toothless]

I have dumb question for OP:

Did you plugged 6 pin power cable from PSU to graphics card before turning on pc?

It won't turn on the GPU if that happened.

 

[Laki89]

It won't turn on the GPU if that happened.

Long time ago i saw some graphics cards performing pretty badly without external power (just motherboard power) but still working and running games
He got somewhere 60% of score that he should get with GTX 1060
Card should spend 120 watt (more or less), if you cut 40% you get 72 watt - which is more or less power that comes from motherboard