Jump to content

Search the Community

Showing results for tags 'benchmark'.

  • Search By Tags

    Type tags separated by commas.
  • Search By Author

Content Type


  • Forum Introduction & Feedback
    • Site Announcements, Rules & New Member Intros
    • TechInferno Forum Feedback
  • Tech News & Reviews
    • News & Reviews
  • Notebook Discussion
    • General Notebook Discussions
    • Notebook OEM Subforums
    • What Notebook Best Fits My Needs?
    • Legacy Articles
  • Desktop & General Hardware
    • General Desktops Discussion
    • Desktop Hardware
    • Overclocking, Cooling & Build Logs
  • Software, Networking & Gaming
    • PC & Console Gaming
    • Video Driver Releases & Discussion
    • Networking
    • General Software Discussion
  • Everything Else
    • Off Topic
  • Legacy Section (Not Actively Supported)
    • DIY e-GPU Projects


  • SVL7 & Klem VBIOS Mods
    • AMD
    • Alienware M11x R3
    • Alienware M14x R2
    • Alienware M17x R4
    • Alienware M18x R1
    • Alienware M18x R2
    • Kepler VBIOS
    • Lenovo Y400-Y500
    • Lenovo Y410p-Y510p
    • Lenovo Y580-Y480
    • Legacy BIOS/VBIOS
    • Maxwell VBIOS
    • Sony Vaio SVS13 / SVS15 series
    • Clevo
  • Utilities

Find results in...

Find results that contain...

Date Created

  • Start


Last Updated

  • Start


Filter by number of...


  • Start






Website URL







Found 9 results

  1. We all love drama between technology companies and the latest deliciousness comes courtesy of AMD. They have a 4 minute video on YouTube that questions whether or not SYSmark is a reliable, objective benchmark to use when evaluating performance between different systems. AMD's video opens with John Hampton, Director Computer Client Products, talking about how it is very important in choosing the right benchmark when evaluating a purchase because if you choose the wrong one , you can end up overpaying for technology or getting a lesser performing PC. To emphasize his point, he indirectly mentions Volkswagen's diesel debacle and uses the latter case as an example of how even established organizations can be misleading. We are then introduced to Tony Salinas, an AMD Engineering Manager, who runs through a couple tests in SYSmark using two notebooks, one that has an Intel i5 CPU and the other with an AMD FX CPU. SYSmark returns a score of 987 for the Intel system and 659 for the AMD one which is a delta of 50%. AMD notes that this is misleading and not reflective of real life performance. To prove this, they run a different benchmark called PCMark 8 on both systems and the results show the Intel i5 system scoring 4199 and the AMD equipped one 3908 which is a delta of 7%. AMD says this is a far more realistic indicator of true performance as PCMark 8 tests the CPU, GPU and video subcomponents rather than emphasizing the CPU like SYSmark does. And just to drive the point home some more, AMD created their own in-house test using Microsoft Office Suite where they run a custom script on both notebooks and then measure the start and end of a task with a time stamp to gauge performance. Their in-house test shows the Intel system completing the task in approximately 61 seconds while the AMD based one finished it in roughly 64 seconds which is a delta of 6-7%. As AMD's results line up with those found with PCMark 8, they contend that PCMark 8 is a more fair and balanced indicator of overall performance while SYSmark is not and therefore should be discarded by any potential customers as a tool for evaluating a purchase. Intel system using AMD in-house test: Test completed in ~61 seconds. AMD system using AMD in-house test: Test completed in ~64 seconds. They also take the time to remind us that the FTC has required SYSmark published benchmarks to contain fine print that notes their benchmark may only be optimized for Intel. Of course, this may have something to do with the fact that AMD resigned from BAPco consortium. Finally, AMD notes that a transparent benchmark should meet very specific criteria which includes being realistic, unbiased, objective and transparent, all of which SYSmark is not apparently. While it's fun to see these types of videos from AMD, they also highlight the fact that the company is facing an uphill battle on all fronts, whether it's CPU or GPU, and 2016 may be the year that makes or breaks them. View full article
  2. Hi guys, I succesfully flashed v2 vbios from prema and I tried 1v and and oc but I couldnt. I think it was power suply. Mine is 230w. And I tried 0.75 0.50 0.25 and 0 volt. But now I m trying game, benchmark screen goes dark and after beeping at 0 volt. Is this a hardware damage? I used just 1v max.(temperature max 70c) I tried cmos reset DDU Xtu uninstall And come back to stock vbios I think my card is dying Please help me!
  3. It’s about time we share first benchmarks of GeForce GTX 1080, new flagship graphics card that will be unveiled tomorrow by NVIDIA. NVIDIA GeForce GTX 1080, 8GB RAM GDDR5X at 10 GHz I think the most interesting detail about GTX 1080 is the new memory. NVIDIA has finally used 8GB ram for its flagship card. It’s no longer an exclusive to mobile solutions. Additionally, the GDDR5X modules are clocked at 2500 MHz (which is shown as 5000+ MHz in 3DMark). However the effective clock is 10000 MHz, which means the bandwidth is somewhere around 320 GB/s (assuming it’s 256-bit wide). On the GPU side we have a huge improvement in terms of frequencies. It is said that GTX 1080 can boost up to 1.8 GHz, with base clock around 1.6 GHz. At the time of writing we are not able to confirm the exact reference clock. For such reason I decided to avoid making comparison charts, so this post will essentially tell you what GTX 1080 is capable of and nothing more. NVIDIA Pascal Series (To be confirmed) May 5th 2016 NVIDIA GeForce GTX 1080 NVIDIA GeForce GTX 1070 NVIDIA Tesla P100 GPU 16nm FF GP104-400 16nm FF GP104-200 16nm FF GP100-890 CUDA Cores ? ? 3584 Memory Type 8GB GDDR5X 8GB GDDR5 16 GB HBM2 Base Clock ? ? 1328 MHz Memory Clock 2500 MHz 2000 MHz 352 MHz Effective Memory Clock 10000 MHz 8000 MHz 1408 MHz Memory Bus 256-bit 256-bit 4096-bit Memory Bandwidth 320 GB/s 256 GB/s 720 GB/s NVIDIA GeForce GTX 1080 8GB — 3DMark11 Performance Like always, we are only looking at graphics score. 3DMark11 Performance preset is rendered at 1280×720 resolution. GeForce GTX 1080 scores 27683 points, which is still above overclocked GM200 cards (~23-25k). Worth noting 3DMark11 is not showing correct GPU clocks, however the new driver already supports GTX 1080 by its name. http://videocardz.com/59558/nvidia-geforce-gtx-1080-3dmark-benchmarks
  4. Desktop, laptop whatever you got in whatever configuration, no holds barred! Mine were done with high quality selected in NVCP, I'd probably squeeze a bit more out of performance textures:
  5. Grab it here: Download Unigine 'Valley' Benchmark 1.0 | techPowerUp It's from the same company that made the Heaven benchmark but this one takes place in a valley.
  6. Well we've done it, the M18x-R1 just crossed the 9k GPU barrier with a SINGLE video card! Major props to @svl7 for his vbios, to nVidia for building Kepler (what an amazing GPU) and to Alienware for designing some awesome cooling, without any of the aforementioned this wouldn't be possible.
  7. So the game has some pretty sweet graphics, lets see how our systems fare. First up is my M18x with the specs + clocks in my sig. Requirements: 1. Keep control panel (e.g. CCC or NVCP) at default quality settings. 2. Use 1920x1080 px resolution. 3. Try to use the latest driver available (Catalyst 12.11 beta 8 + Geforce 310.64 beta). If you prefer using WHQL you may. You may OC as high as your system will allow and post your results. Game settings Ultra Preset (everything turned up all the way and all checkboxes ticked) + 8X MSAA: Ultra Preset + 4xMSAA: High Preset + 4xMSAA: We will look at the avg + min fps since the max can vary from test to test.
  8. Fraps video test with rage im posting cause when im looking for a laptop to purchase i really enjoy these to see where the laptop stands with current gaming http://www.youtube.com/watch?v=SWZ-WAY27CQ
  9. *In response to a PM from DR650SE for overlocking advice… figured everyone can discuss this togetherso I posted it instead. "You can followanything I wrote as my settings by all my bench pics but here's anything thatcomes to mind. I had to get high cpu scores on my r2 when I had it to keep upon the lack of having ever changing factory pasting on the gpus and no dice orac so I thought I was fuxxed so I found my own way without those methods. I mayof had lower gpu scores but had my cpu at 8kcpu on 920xm and on 3dmark11 ispretty good right for the R2? I am the beginner Ilearned from you guys and from back when on Wikia free forums when it beforeT|I was up. I learned how to flash vbios from you back then only. So the theory. Youknow this stuff but having it all work together is the key. I had 5870m not6970m in the r2 so there's emphasis on YOU using every method possible forsaving power and keeping cool. Maybe you and Stam really should work togetherto make your supply dual parallel power supplies and I'll do it on my m18x withyou guys... R2 would have 480watts for play and 660watts for m18x. Okay now thetechniques since I use the normal power supply like you. You guys had amazingdiscussions on it so far. I am also running into power shutdowns like when over115amp and watts are set high on cpu... that's happening without my gpusoverclocked. Some of this you'vesaid yourself DR650SE. Try sticking to vbios overclock not using ULPS off inwindows that allows power saved while idling before the test and there would beless heat already for the second gpu... So don't disable ULPS just make like 20variations of your overclock in organized fashion on a usb stick with RBE thenyou can quickly move your next overlock vbios files in right place and flashand check settings quicker. Before your main run you can still use ULPS off andTrixx to find your range of good clocks then turn ulps back on and flash yourplanned overclock. Turn off all bios options for Bluetooth and PCI card slotsremove all USB stuff, use just a single or dual ssd drive if in raid 0 anddon't have a motor based platter hard drive while benching. Keep the powersupply cool also, see my makeshift cooler on the website. Cool components equalless required watts, MHz, voltage id, amperage etc for the same performance orbetter and heat raises these as heat goes up. So we need all ways possible tolower wattage coming from your power supply compare numbers on my highest cpubenches in the r2 bench thread to see which voltage I used, watts and clockssince they can help. Example 4ghz mayscore less than an undervolted 3.75ghz. feel the config out with wprime if thewprime scores good on that config probably will be good for 3dmark. So undervolt even ifit sacrifices a slight bit of MHz the scores can be higher from lower temps. SoI'm saying use passive cooling ie undervolt to the point of just stable enoughto survive the test. Amps will shut off system faster than watts ... Tdc andtdp but if you use a little higher amps and limit your watts.. The load canquickly be drawn on demand and allowing high wattage peaks but no crash sincethey peak and go back with cpu thermal envelope. Use throttle stop for morethan just over clock have overclock on one profile and on the other set yourlowest multiplier locked like is it 12 on i7 first gen? Its 8 on Sandybridge.And also on that profile check cmod or one of them and drop it anywhere from25%-87.5% so your chips cores are running well under natural idle loads andtemps will drop further leave the cpu cooling awhile so when you launch thebench you restore your overclock but cpu cooler that's directly contacting thepaste and cpu is also extra cool where as its not if you didn't do that justthe entire cpu heatsink would be cool not the area I mentioned. Now if you look at HWinfo summary I discoveredyou see what actually happens during cmod less than 100%... I noticed your cpucores suddenly operate at the percent speed you set the cmod too but also thememory does this and even the baseclock drops the same percent giving 100%system cooling via passive drops in speed in almost all clocks of the systemallowing components to absorb more of your cooling on the system. Cool rightthat is additional passive cooling since its using lack of voltage and MHz tothe cores to keep them extra frosty before the benching starts. Fans should be maxedon gpus and cpu with HWinfo. Those heat sinks need to be extra cool beforehand. I never benched with the covers ever off but the battery out for sure socharging isn't happening. All Alienware lights out, you will need Alienware OSDenabled set to go dark and no LED's lit even status lights. When windows startsand lights go out fully on keyboard etcclose OSD process manually and start hitting stop on every unnecessary servicein windows services window. Next closeanything remaining in task manager that's unnecessary since some will be foundhere left. Make sure your plugged into outlet with enough juice some in housescan just be appropriate for lighting. Set screen brightness all the way down.Stop most startup apps in ccleaner use disable since it is easy to enable againthis way. You can keep system covers off since your trying to squeezeperformance. Use the dice on everything even the heatsinks. Use your dice potand ICD 24 its amazing paste. Repaste your CPUs and gpus including Vram andother heat creating pieces. Pre chill your system by idling before kicking theoverclock in and starting the bench. If need create a hotkey to lower cpu orraise in certain scenes of the benches. I haven't but it will help. Undervoltundervolt undervolt! I'm honored to help, I am the newbie amongstyou guys. I use a basic notebook cooler they do help you should too since Inotice my aluminum frame helps cool components a giant amount in fact thealuminum always felt cold. You should use air conditioning too or cool yourroom extra for it even if using dice all components should be cool. Use thehighest baseclock possible that's stable. A high baseclock is priority.Undervolt ram unless its the lowest voltage you can get stable. My ram for the8kcpu in 3dmark11 was stock Dell ram I ran high baseclock but was able to keep1.5v ram. It was like 146-149 baselock with fsb on r2. (m18x cant use fsb). Theram I had used 1.5v with a huge over clock from its stock 1333mhz. Unlike myHyperX I bought for my M18x at 1600mhz which are 1333mhz chips at 1600mhztimings but require 1.65 volts. I wish they would run at lower volts, for me ifI overlock on my baseclock on my m18x my max ram volts is 1.65 in bios is so ithandles to a certain point and would have benefited from 1.7v for extremeoverclocks... But for your r2 try 1.5volt on ram to help heat too and powertoo. If you have Dell ram handy try that you never know one may handleoverlocks and undervolting better. Use a thick lead from your powersupply towall that's rated at high amps. The M18x lead is thickest I've seen. Setwindows to performance in the computer properties so there's no excessivesystem animations etc. Make sure your drive or drives are defragmented and ifits SSD look into a defrag program that aids in SSD i/o operations withouttraditional defrag. In the past I had used Diskeeper or Perfect Disk fordefragging... there a few companies that use similar invisible defragging thathave no impact on user operations while the system is in use. Also ramdisks forpagefile or the app your using could help too. Another is called Fancy Cachethat I think Unreal posted just do a search for it. There's been a thread aboutthose types of programs haven't tried it but there's guaranteed a benefit sinceram transfer rates are much higher than SSD and regular drives are just a jokecomparatively. The higher your Voltage ID at the given moment the more tdc andtdp it will require. My same settings set with only difference being VID boostrequire me too raise tdp and tdc a bit to maintain the same speed. Cheers!Happy overclocking. If you have any unmentioned techniques I'm all ears too. *This is just some basic rambling and observations… so just some thoughts I am in no way sayingany of this is concrete. Feel free to discuss these and any other methods youhave used. The only required proccesses if not using internet or graphics just leave your companies graphic services running if needed this is after stopping uneeded services just close the remaining uneeded processes http://i1237.photobucket.com/albums/ff474/marcwwaters86/Theonlyrequiredproccessesifnotusinginternetorgraphicsjustleaveyourcompaniesgraphicservicesrunningifneededthisisafterstoppinguneededservicesj.png The only required services when not using graphics or internet for graphics just leave your graphics companies services running http://i1237.photobucket.com/albums/ff474/marcwwaters86/Theonlyrequiredserviceswhennotusinggraphicsorinternetforgraphicsjustleaveyourgraphicscompaniesservicesrunning.png
  • Create New...

Important Information

By using this site, you agree to our Terms of Use. We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue.