Posted March 25, 2014 Actually, your screenshot shows fairly good that the CPU is in fact 'bottlenecking' the GPUs. I come to that conclusion, because the first core is at its limit for almost the whole time the image shows. You could have your cpu running at 100% and your gpu's at 99%, that's a very rare case to see and another reason why you should ignore the cpu load. So if your cpu is hitting its limit doesnt mean its bottlenecking a gpu. In the recent WAN-Show Linus mentioned that DirectX is programmed for single core operation and therefore can't utilize multiple cores at the same time. Hmm? As I showed the core loads, you see in total 37 threads and I'll be predicting that this engine has one big major thread and one smaller as you could see in my screenshot and the other ones are minor that could easily move to the 2nd major thread without any performance loss. As you know you can only execute a single thread on one core so you kinda want to split the main threads and that should allow you to take performance advantages of the extra cores. Unigine is just another shitty multithreaded engine. A trick to improve performance could be to disable Intel's Hyperthreading Technology, but I am not sure about that. As far as I know, Hyperthreading allows two cores to share one cache or rather one core to split up into two seperate with one cache. Hyperthreading just allows you to get more out of a physical core. To give an example, two instructions one being integer and the other being floating point and lets say the first instruction goes in your ALU (integer core) and your 2nd instruction needs to wait for the 2nd cycle that would mean something like 5+2 + 5/3 is done in 2 cycles, so Intel thought: "why not have them both executing at in one cycle instead of two?" thats why HT is invented.Requires atleast 2 threads otherwise HT wont work. I think your explanation of HT is from Bulldozer/vishera They share their caches, not intel. A trick to improve performance could be to disable Intel's Hyperthreading Technology, but I am not sure about that. As far as I know, Hyperthreading allows two cores to share one cache or rather one core to split up into two seperate with one cache. If that is the case, my tip could work. I would definitely give it a try if your mainboard allows for it. Please, let me know the results if you do try it out. Tried this plenty of times, well not really. I also want to retest tomorrow. That time with the same overall framerate, assuming that then the GPU load is more even between the two benchmarks. (I have mistaken the scores of my previous test, too. The Valley score is actually higher than the Heaven score. I think this is caused by the tessellation.) It's a bit flimsy to compare the scores though because it seems to be only one loop is cpu bound. If you want to do it, just have both of your gpu's at stock speed and both 100% fan speed to avoid random throttling that gpu boost 2.0 causes, try one run at a low cpu clock of 3.5GHz and 4.8GHz orsomething. Youre better off comparing the frames at that scene (6 I think it was) where I was cpu bound. You could try makes it a bit easier how much an OC can effect your gpu loads: http://i.imgur.com/vgwKIE5.jpg The moment you're looking at the buildings the gpu's jump up to 99% or you get the fpscap of 200, tried earlier doing 3.5GHz with 6 cores vs 4.8GHz 4 cores no HT was a gain of 20-25fps. GPU loads went from 30/30 to 50/50%. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 780ti SC out the box...at 1440p Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 I'm gonna build a system with R9 290 Sapphire Tri-X. That test bench will be interesting, I'll be really pushing it to the limit. FANBOY OF: PowerColor, be quiet!, Transcend, G.Skill, Phanteks FORMERLY FANBOY OF: A-Data, Corsair, Nvidia DEVELOPING FANBOY OF: AMD (GPUS), Intel (CPUs), ASRock Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 I'm gonna build a system with R9 290 Sapphire Tri-X. That test bench will be interesting, I'll be really pushing it to the limit. Get the powercolor PCS+ if you can, ive found it to be superior to the tri-x Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 Get the powercolor PCS+ if you can, ive found it to be superior to the tri-xWhat sort of difference do you get between the two? I'll try to get it if I can find one as cheap as the Sapphire atm FANBOY OF: PowerColor, be quiet!, Transcend, G.Skill, Phanteks FORMERLY FANBOY OF: A-Data, Corsair, Nvidia DEVELOPING FANBOY OF: AMD (GPUS), Intel (CPUs), ASRock Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 I has been a while since this has been updated. I posted my scores 10 days ago and they still aren't on the spreadsheet. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 What sort of difference do you get between the two? I'll try to get it if I can find one as cheap as the Sapphire atm check out my channel, I put up a comparison video between the two Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 26, 2014 check out my channel, I put up a comparison video between the two Thanks, been looking for a review between the two FOREVER. FANBOY OF: PowerColor, be quiet!, Transcend, G.Skill, Phanteks FORMERLY FANBOY OF: A-Data, Corsair, Nvidia DEVELOPING FANBOY OF: AMD (GPUS), Intel (CPUs), ASRock Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 28, 2014 make also a top 30 for single gpu instead of mixing CF configs with single solutions Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 28, 2014 make also a top 30 for single gpu instead of mixing CF configs with single solutions There is one, look at the top of the list for each GPU setup. -The Bellerophon- Obsidian 550D-i5-3570k@4.5Ghz -Asus Sabertooth Z77-16GB Corsair Dominator Platinum 1866Mhz-x2 EVGA GTX 760 Dual FTW 4GB-Creative Sound Blaster XF-i Titanium-OCZ Vertex Plus 120GB-Seagate Barracuda 2TB- https://linustechtips.com/main/topic/60154-the-not-really-a-build-log-build-log/ Twofold http://linustechtips.com/main/topic/121043-twofold-a-dual-itx-system/ How great is EVGA? http://linustechtips.com/main/topic/110662-evga-how-great-are-they/#entry1478299 Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 28, 2014 I has been a while since this has been updated. I posted my scores 10 days ago and they still aren't on the spreadsheet. Looked at earlier posts and it looks like last update was 15th of February. Hive Fire - Build of the Month March 2014. Simple 760t Mod - erly ideas for Corsair 760t Mod Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 28, 2014 OP havnt been online since Mar 16 2014 07:46 AM, so he havnt seen the posts after that. CPU i7-930 @ 4GHz Mobo Asus P6X58D-E GPU Gigabyte 970 G1 RAM Patriot Division 4 Viper Extreme 8GB HDD Western Digital 1TB SSD Samsung 840 pro 128GB PSU Corsair TX 750w Cooling Corsair H100i Os Win 10 Pro Monitor AOC G2460PG Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 28, 2014 OP havnt been online since Mar 16 2014 07:46 AM, so he havnt seen the posts after that. Funny, thats the day I posted. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 28, 2014 I've been lazy in regards to updating the list, mainly due to work/social life. I'll get it done today, i'll be looking for someone to help update the list throughout the week. If you're keen shoot me a pm and when i get back on i'll go through them This was his last post in this thread; it looks like he is getting too swamped to update this regularly and didn't get/accept any offers to help update the list. System CPU : Ryzen 9 5950 doing whatever PBO lets it. Motherboard : Asus B550 Wifi II RAM 80GB 3600 CL 18 2x 32GB 2x 8GB GPUs Vega 56 & Tesla M40 Corsair 4000D Storage: many and varied small (512GB-1TB) SSD + 5TB WD Green PSU 1000W EVGA GOLD You can trust me, I'm from the Internet. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 29, 2014 Intel 4670k GTX 660ti @ 1228Mhz / 1674Mhz 37.2 FPS 1556 Score http://imgur.com/Xm38eui Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 30, 2014 I'm sure he'll get around to updating eventually. If you're really wanting to get onto a chart that bad you can always post in the OCN Valley thread which is probably one of the most active Valley threads on the web. I'll warn you though the scores over there are mental! http://www.overclock.net/t/1360884/official-top-30-unigine-valley-benchmark-1-0/0_50 Intel Core i7 4930K @ 4.7GHz | Asus Rampage IV Extreme | 2 x EVGA GTX Titan SC (1254MHz) | 16GB Patriot Viper Extreme DDR3 2133MHz (4 x 4GB) | Corsair AX1200 | Silverstone Temjin TJ11 | Corsair Force 3 240GB (System) | 2 x Intel 320 160GB SSD (Dedicated Gaming Drives) | Hitachi Deskstar 1TB (Data) | MS Windows 10 Pro | EK Supreme HF/FC-Titan/Rampage IV Extreme blocks | Hardware Labs GTX 560/240 rads | Alphacool VPP-655 D5 pump | Bitspower mod kit/pump top/fittings/120mm res Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 30, 2014 I'm sure he'll get around to updating eventually. If you're really wanting to get onto a chart that bad you can always post in the OCN Valley thread which is probably one of the most active Valley threads on the web. I'll warn you though the scores over there are mental! http://www.overclock.net/t/1360884/official-top-30-unigine-valley-benchmark-1-0/0_50 On that one, 77th. Yes! Yeah high scores for sure. Nope not wanting to get on a chart that bad, just bringing it to everyone's attention that the list inst being updated. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 30, 2014 . Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 30, 2014 Mobo: Asus M5A97 R2.0 Cpu: AMD FX 8350 (4.2Ghz - 4.4Ghz Turbo) Gpu: Gainward Geforce gtx 770 Phantom 4gb (Stock) Ram: 8gb (2x4) Corsair XMS3 1600Mhz Nvidia 3D settings. Extreme HD + Extreme HD PhysX Surround Just to take an example of the difference between having the Nvidia gpu settings i posted a screenshot of above at performance and by having them at absolute max quality (like i do), would be: Skyrim: With max quality on the gpu settings and the game settings at highest = 60-90 fps. With gpu settings set to performance and game settings to highest = 120-210 fps. So i'm guessing that this do affect my benchmarking results quiet a bit. However, i always want the best graphics i can get, so just felt natural to me to keep the settings at quality instead of performance even during these test runs. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 31, 2014 [...] Retried with the following settings: Heaven: Extreme, not tessellation, 1920x1080 (fullscreen) Valley: ExtremeHD, 1920x1080 (fullscreen) Intel Core i5-4670k @ stock Palit GTX 770 @ stock Intel Core i5-4670k @ 2,9GHz Palit GTX 770 @ stock Result CPU-load overall stays about the same and is roughly equal between the two benchmarks. When the CPU gets undeclocked to 2,9GHz: - GPU-load and clocks stay about the same. - GPU-power ist decreased by about 10%. - The FPS in Heaven are decreased by 43,28%. - The FPS in Valley are decreased by 41,05%. Conclusion Once both benchmarks are set up to achieve about equal scores, the load on the CPU and GPU are nearly even between the two. The CPU-performance makes a significant difference below a system-specific performance. In my case however, overclocking the CPU only makes a neglectable difference. That might change, once another card is added. Reduce > Reuse > Recycle Build-log (way out of date) Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 31, 2014 Retried with the following settings: Like I said the difference will be minor with a single gpu, not enough headroom. So a good 10% difference, would have been more if all scenes were cpu bound; only 2 scenes were cpu limited all others were gpu bound. Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 31, 2014 Syfes Intel i7-2600k GTX 780 Ti @ 1489Mhz / 1849Mhz 82.2 FPS 3441 Score Desktop: Core i7-9700K @ 5.1GHz all-core = ASRock Z390 Taichi Ultimate = 16GB HyperX Predator DDR4 @ 3600MHz = Asus ROG Strix 3060ti (non LHR) = Samsung 970 EVO 500GB M.2 SSD = ASUS PG279Q Notebook: Clevo P651RG-G = Core i7 6820HK = 16GB HyperX Impact DDR4 2133MHz = GTX 980M = 1080p IPS G-Sync = Samsung SM951 256GB M.2 SSD + Samsung 850 Pro 256GB SSD Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 31, 2014 Author Syfes Intel i7-2600k GTX 780 Ti @ 1489Mhz / 1849Mhz 82.2 FPS 3441 Score what is the core speed in gpuz, valley is probably wrong? Rig Specs: AMD Threadripper 5990WX@4.8Ghz Asus Zenith III Extreme Asrock OC Formula 7970XTX Quadfire G.Skill Ripheartout X OC 7000Mhz C28 DDR5 4X16GB Super Flower Power Leadex 2000W Psu's X2 Harrynowl's 775/771 OC and mod guide: http://linustechtips.com/main/topic/232325-lga775-core2duo-core2quad-overclocking-guide/ http://linustechtips.com/main/topic/365998-mod-lga771-to-lga775-cpu-modification-tutorial/ ProKoN haswell/DC OC guide: http://linustechtips.com/main/topic/41234-intel-haswell-4670k-4770k-overclocking-guide/ "desperate for just a bit more money to watercool, the titan x would be thankful" Carter -2016 Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted March 31, 2014 what is the core speed in gpuz, valley is probably wrong? Hey you're right, it's off by 170Mhz, it scales just fine though. There's just always a 170Mhz offset in valley. This is an EVGA GTX 780 Ti Classified K|ngp|n edition using stock aircooler, correct frequencies below: Syfes Intel i7-2600k GTX 780 Ti @ 1319Mhz / 1849Mhz 82.2 FPS 3441 Score Desktop: Core i7-9700K @ 5.1GHz all-core = ASRock Z390 Taichi Ultimate = 16GB HyperX Predator DDR4 @ 3600MHz = Asus ROG Strix 3060ti (non LHR) = Samsung 970 EVO 500GB M.2 SSD = ASUS PG279Q Notebook: Clevo P651RG-G = Core i7 6820HK = 16GB HyperX Impact DDR4 2133MHz = GTX 980M = 1080p IPS G-Sync = Samsung SM951 256GB M.2 SSD + Samsung 850 Pro 256GB SSD Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Posted April 1, 2014 Those 0,3 FPS and 13 points make all the difference. Karikiro Intel Core i5-4670k @ 4,2 GHz (funcore = 3,9 GHz) Nvidia GTX 770 @ 1333 MHz / 2000 MHz (4001 MHz) 54,8 FPS 2292 Score Reduce > Reuse > Recycle Build-log (way out of date) Link to comment Share on other sites More sharing options... Link to post Share on other sites More sharing options...
Create an account or sign in to comment
You need to be a member in order to leave a comment
Create an account
Sign up for a new account in our community. It's easy!
Register a new accountSign in
Already have an account? Sign in here.
Sign In Now