Extremely low Cinebench 15 score (74 cb) on i5-4590

Clyder,

That's a strange result so low it might be anomalous related to the nature of the test. Cinebench evaluates the system's capabilities to run Maxon animation software, and a CPU that is not hyperthreading may not perform well. The other Cinebench test component is OpenGL and if you're running a DirectX- oriented gaming GPU, that may also tip the score.

What is the experience of using it?

My sense is that Cinebench may not be showing good results because of the way it tests what it looks for, if that makes any sense.

My suggestion is to, in effect, test the test. Download and run Passmark Performance Test- there's a 30-day free trial- and post the results back here.

I use both Passmark and Cinebench to evaluate my workstations. which use 3D CAD and rendering software that are also OpenGL and beenfit from thread distibution,

My analysis / simulation / rendering system:

HP z620 (2012) (Rev 3) 2X Xeon E5-2690 (8-core @ 2.9 / 3.8GHz) / 64GB DDR3-1600 ECC reg) / Quadro K2200 (4GB) + Tesla M2090 (6GB) / HP Z Turbo Drive (256GB) + Samsung 850 Evo 250GB + Seagate Constellation ES.3 (1TB) / Creative Sound Blaster X-Fi Titanium PCIe sound card / 800W / Windows 7 Professional 64-bit > Logitech z313 2.1 speakers > HP 2711x (27" 1980 X 1080)
[ Passmark System Rating= 5675 / CPU= 22625 / 2D= 815 / 3D = 3580 / Mem = 2522 / Disk = 12640 ] 9.25.16
[ Cinebench R15: OpenGL= 119.23 fps / CPU = 2209 cb / Single core 130 cb / MP Ratio 16.84x] 10.31.16

The Passmark and Cinebench tests in this case agree on a strong CPU performance because they're looking for similar individual parameters. The 16 cores /32 threads at up to 3.8GHz produces a high CPU score in both Passmark and Cinebench because of the way the tests are weighted towards what the makers establish as a experience of a fast system.

If I ran the HP z620 on a test strongly weighted towards gaming FPS, the i5-4590 would score better than the E5-2690 since the i5-4590 has a Passmark single-thread rating of 2112 and the Xeon E5-2690 rating is 1873. In games- and in fact most kinds of software except CPU rendering and some custom algorithm scienticc applications such as Matlab and Wolfram, the single-thread rating is more critical than the number of calculation cycles- which the 16 cores and 32 threads of the E5-2690 do very well: 22625 to the i5-4590 average of 7182.

But, the Passmark Test scores the Quadro K2200 (4GB) + Tesla M2090 (6GB) GPU combination the same as a single K2200 while the Ocatanebench test can rate the Tesla GPU coprocessor as well and scores that comibations (that cost $350 + $86) similarly to a Quadro M5000 8GB- a $1,900 GPU.

There's speed and then there's the other kind of speed.

Cheers,

BambiBoom

Clyder,

The new information is a help to narrow down the problem. In the absence of any error messages, if the Passmark 3D Mark for the GTX 960 is within or very close to the average- it seems to be working, but the CPU score is very poor, here are some speculations:

1. That there is something amiss in the BIOS, especially as regards to enabling all processors and all cores.

2. Check that BIOS is the latest version.

3. It's worthwhile at this point to review the BIOS settings in detail.

4. check the state of hyperthreading. The i5-4590 is not a hyperthreading CPU and if hyperthreading is set to ON, it may cause a conflict.

5. There is also the matter of the power option setting, which should be set to "performance".

6. If 1-5 don't help, if you're confident, clear CMOS which will reset everything to default.

At least these steps will eliminate some of the easy items.

If nothing in 1>6 helps:

For reference I'm curious:

1. what is the model of the motherboard

2. as to how long you've had this system,

3. how long the performance has been poor,

4. have there been any recent changes,

5. and have there been any previous tests?

Cheers,

BambiBoom

Clyder,

Sorry for the late reply, all last evening I was "banned" from the forums. I was going to suggest trying a previous version of the BIOS.

For reference: Checking Passmark, there are 22 systems using an i5-4590 on the ASRock H97 Anniversary. The top CPU score is 8033 and the lowest is 6393. All of the results are @ 3.3GHz base speed and 3.7GHz Turbo.

The next items to consider:

1. Run the Intel Processor Diagnostic Tool

2. Although none of the systems on Passmark appear to be oveclocked, it's still worth verifying the CPU temperatures.

3. Update the chipset driver

4. Run memtest

5. Thorough malware check

Do you another LGA1150 CPU around?

We'll hope it's the easy thing.

Cheers,

BambiBoom

Clyder,

You may well have seen the ASRock H97 Anniversary driver page.
The other possibilities that I can think is a corrupted registry but that is not a satisfying explanation as to why the CPU and memory diagnostics are passing plus the temperatures are normal. It's as though everything is correct and working, but performance is poor. I don't understand how the CPU can be running so far below normal and still pass all those tests plus the lack of error messages. The lack of error messages was the reason that seemed to point to a BIOS problem.

If the system was working well and then changed, the key has to be some event or alteration at that point in time. Do you have any system restore points dated before the problem?

If the BIOS is fine, settings are correct and temperatures within tolerance, the final step might be to consider a careful double-checking of all hardware and each connection, check that the internal speaker is working too in case error codes are being missed, format the drives, and reinstall Windows, graphics driver, and the software, but my thinking is now gravitating towards a gradual motherboard failure. It's a bit of a fuss, but just to eliminate it as a possibility, it may be worth taking out the CPU and inspecting for burnt capacitors and check the socket for bent pins.

But, I can't get past the system running properly earlier and current lack of error codes.

This is a real puzzler.

Cheers,

BambiBoom