*Official* NBR Desktop Overclocker's Lounge [laptop owners welcome, too]

Thanks Brother Fox.

 

I've added some more analysis to my massive post above, mainly gaining perspective on the results I've seen, it's under EDIT#4.

 

That's a bit much considering that AMD doesn't lock its processors and (particularly on MT workloads) blows everything else out of the water shy of the 9980XE. Why do you care about the clock frequency if the wall clock performance is there?

AMD does a lot of work to optimize performance on the fly, what with PBO and the X570 giving the CPU information about the VRM's (which doesn't seem to have much effect -- there's little if any difference in CPU performance between the X570 and X470 or even B450). The fact that you can't do better either means that you need to up your game or that AMD really is squeezing everything possible out of the CPU rather than leaving something on the table for only a few people with m4d sk1llz to extract.

The fact that AMD is matching if not beating the 9900K with an 8 core chip while thrashing it with 12 cores, and is getting everything it can out of the chips, sure looks to me like it's doing whatever it takes to win. Leaving performance on the table would be doing just the opposite.

 

Individual screen images: https://drive.google.com/file/d/13CQoFP36Kc8KXQAQAh348_bZStPdvcwG/view?usp=sharing

 

Stock performance is nice, but it is not that important to me. I do not see any AMD CPUs that overclock well enough to keep up with my overclocked 7960X. What matters most to me is how much I can extract from it by pushing it harder and harder. It does not seem like AMD makes products that do well with that. I do not want to buy a bellybutton CPU that runs the essentially same for everyone that owns one and does not respond well to overclocking efforts. To me that is too boring and I have no interest in owning it. I am not saying they perform poorly or that they are junk. They are not. Just too boring for me to waste my money on one since they don't overclock worth a damn. Because they do not, they are incapable of meeting my expectations even though they do provide very nice stock performance.

 

Using Throttle Stop and HWiNFO for CPU wattage, I added TGP + CHP + CPU wattage and it came to exactly what HWiNFO was displaying for Total System Power in-game. Pleased to see accuracy there at least. I'm still reading the User Guide, will use while gaming for a few days to see how it goes.

 

Interesting. Don't know if this vBIOS is going to work well or not, but if the wattage is accurate it might be awesome.

 

I look forward to hearing you results. I might bust out the programmer and try and force flash it later. But knowing that it does higher voltage and V/F curve unlocked it might be the one we've been wanting.

 

Hey Brother thanks so much. Travel mercies going up for you tomorrow.

 

It is a really good vBIOS. I haven't been able to beat any of my benchmark high scores with the K|INGP|N vBIOS, but it's right there very close. And, yes... having the voltage curve unlocked with higher voltage is a lot better than stock vBIOS. Even if I cannot beat my current high scores, it is better for that reason alone. It's so close that it is within a normal margin of error. And, I can bench it with the core at 2250-2265 without crashing now (but the score isn't any higher than 2235). Really liking that 2000W power limit.

https://www.3dmark.com/compare/fs/19853616/fs/19466211#

Thanks, bro. I am really excited about going on vacation.

Be back next weekend guys. Probably won't be on the forum at all next week.

 

Doomed to fail... http://forum.notebookreview.com/thr...l-breaks-intel-xtu-and-doesnt-install.829667/

 

Took me busting out the test clip and usb programmer and I got her to take the flash. This GPU is unleashed now with the daily driver functionality I was looking for.

 

TGP value reported in NVidia FrameTime is the same figure that's displayed in GPUz and HWInfo for GPU Power, which is the power consumption in Watts of the whole GPU card.

Wow, that's insane isn't it, nearly 4 times the amount I've ever put through my GTX 1070! (264W in Furmark)

That PowerLimit.bat, which GPUs is that compatible with? I see you're using a Galaxy vBIOS at the moment, didn't you test a Galaxy HOF vBIOS before and you found it to be rubbish?

 

Maybe because I am on mobile/laptop, but my RTX 2070 offers no wattage information via GPU-Z. It might be in HWiNFO but the ease of FrameView display is nice!

 

I'm using an Asus Xonar U7 external USB soundcard, and that provides great sound to my AKG K702 headphones, way better than the onboard audio of my motherboard which sounds muddy in comparison. I think there's a new version of the Xonar U7 that's come out over the last year or two, I bought mine in 2013 I think. Depends what you want the soundcard for, but I can easily recommend Xonar U7 for my uses - just to power my headphones.

 

Not the same.

 

This can vary by vBIOS and can also vary by driver version. When testing vBIOS mods with @Prema this data output to GPU-Z, including the power (W) and TGP % would disappear and reappear depending on the firmware version present on the GPU. When it disappears in GPU-Z it often disappears in other monitoring software since the firmware and/or driver are not providing any data for the software to display. The same is true of the iCX data output on my EVGA cards. The sensors are there and you can always see them using EVGA Precision XOC software even when the vBIOS doesn't support the data output to third party software like GPU-Z, AIDA64 and HWiNFO64. The EVGA Precision XOC software finds and reads the data directly from the sensors without the underlying firmware support. For example, if I flash Aorus Extreme vBIOS on my 2080 Ti, the iCX data can only be seen with Precision XOC. When I use the stock EVGA vBIOS it shows in all software, unless I use a driver version that has a bug that doesn't read it.

Yes, it was rubbish. Still is totally rubbish. Different vBIOS. The only vBIOSes I have tested that are not totally rubbish are the stock EVGA firmware, Aorus Extreme, K|INGP|N and this one.

I am only using the batch file to simplify entering the command syntax. The command that it is running is for Nvidia SMI. So the ability to use the command is part of the driver package. Not every GPU will support it. For example the Quadro K1100M card in my laptop does not. Its power limit is locked, not adjustable.

Code:

C:\Program Files\NVIDIA Corporation\NVSMI>nvidia-smi.exe -h
NVIDIA System Management Interface -- v425.51

NVSMI provides monitoring information for Tesla and select Quadro devices.
The data is presented in either a plain text or an XML format, via stdout or a f
ile.
NVSMI also provides several management operations for changing the device state.


Note that the functionality of NVSMI is exposed through the NVML C-based
library. See the NVIDIA developer website for more information about NVML.
Python wrappers to NVML are also available.  The output of NVSMI is
not guaranteed to be backwards compatible; NVML and the bindings are backwards
compatible.

http://developer.nvidia.com/nvidia-management-library-nvml/
http://pypi.python.org/pypi/nvidia-ml-py/
Supported products:
- Full Support
    - All Tesla products, starting with the Kepler architecture
    - All Quadro products, starting with the Kepler architecture
    - All GRID products, starting with the Kepler architecture
    - GeForce Titan products, starting with the Kepler architecture
- Limited Support
    - All Geforce products, starting with the Kepler architecture
nvidia-smi [OPTION1 [ARG1]] [OPTION2 [ARG2]] ...

    -h,   --help                Print usage information and exit.

  LIST OPTIONS:

    -L,   --list-gpus           Display a list of GPUs connected to the system.

    -B,   --list-blacklist-gpus Display a list of blacklisted GPUs in the system
.

  SUMMARY OPTIONS:

    <no arguments>              Show a summary of GPUs connected to the system.

    [plus any of]

    -i,   --id=                 Target a specific GPU.
    -f,   --filename=           Log to a specified file, rather than to stdout.
    -l,   --loop=               Probe until Ctrl+C at specified second interval.


  QUERY OPTIONS:

    -q,   --query               Display GPU or Unit info.

    [plus any of]

    -u,   --unit                Show unit, rather than GPU, attributes.
    -i,   --id=                 Target a specific GPU or Unit.
    -f,   --filename=           Log to a specified file, rather than to stdout.
    -x,   --xml-format          Produce XML output.
          --dtd                 When showing xml output, embed DTD.
    -d,   --display=            Display only selected information: MEMORY,
                                    UTILIZATION, ECC, TEMPERATURE, POWER, CLOCK,

                                    COMPUTE, PIDS, PERFORMANCE, SUPPORTED_CLOCKS
,
                                    PAGE_RETIREMENT, ACCOUNTING, ENCODER_STATS,
FBC_STATS
                                Flags can be combined with comma e.g. ECC,POWER.

                                Sampling data with max/min/avg is also returned

                                for POWER, UTILIZATION and CLOCK display types.
                                Doesn't work with -u or -x flags.
    -l,   --loop=               Probe until Ctrl+C at specified second interval.


    -lms, --loop-ms=            Probe until Ctrl+C at specified millisecond inte
rval.

  SELECTIVE QUERY OPTIONS:

    Allows the caller to pass an explicit list of properties to query.

    [one of]

    --query-gpu=                Information about GPU.
                                Call --help-query-gpu for more info.
    --query-supported-clocks=   List of supported clocks.
                                Call --help-query-supported-clocks for more info
.
    --query-compute-apps=       List of currently active compute processes.
                                Call --help-query-compute-apps for more info.
    --query-accounted-apps=     List of accounted compute processes.
                                Call --help-query-accounted-apps for more info.
    --query-retired-pages=      List of device memory pages that have been retir
ed.
                                Call --help-query-retired-pages for more info.

    [mandatory]

    --format=                   Comma separated list of format options:
                                  csv - comma separated values (MANDATORY)
                                  noheader - skip the first line with column hea
ders
                                  nounits - don't print units for numerical
                                             values

    [plus any of]

    -i,   --id=                 Target a specific GPU or Unit.
    -f,   --filename=           Log to a specified file, rather than to stdout.
    -l,   --loop=               Probe until Ctrl+C at specified second interval.

    -lms, --loop-ms=            Probe until Ctrl+C at specified millisecond inte
rval.

  DEVICE MODIFICATION OPTIONS:

    [any one of]

    -e,   --ecc-config=         Toggle ECC support: 0/DISABLED, 1/ENABLED
    -p,   --reset-ecc-errors=   Reset ECC error counts: 0/VOLATILE, 1/AGGREGATE
    -c,   --compute-mode=       Set MODE for compute applications:
                                0/DEFAULT, 1/EXCLUSIVE_PROCESS,
                                2/PROHIBITED
    -dm,  --driver-model=       Enable or disable TCC mode: 0/WDDM, 1/TCC
    -fdm, --force-driver-model= Enable or disable TCC mode: 0/WDDM, 1/TCC
                                Ignores the error that display is connected.
          --gom=                Set GPU Operation Mode:
                                    0/ALL_ON, 1/COMPUTE, 2/LOW_DP
    -lgc  --lock-gpu-clocks=    Specifies <minGpuClock,maxGpuClock> clocks as a
                                    pair (e.g. 1500,1500) that defines the range

                                    of desired locked GPU clock speed in MHz.
                                    Setting this will supercede application cloc
ks
                                    and take effect regardless if an app is runn
ing.
                                    Input can also be a singular desired clock v
alue
                                    (e.g. <GpuClockValue>).
    -rgc  --reset-gpu-clocks
                                Resets the Gpu clocks to the default values.
    -ac   --applications-clocks= Specifies <memory,graphics> clocks as a
                                    pair (e.g. 2000,800) that defines GPU's
                                    speed in MHz while running applications on a
 GPU.
    -rac  --reset-applications-clocks
                                Resets the applications clocks to the default va
lues.
    -acp  --applications-clocks-permission=
                                Toggles permission requirements for -ac and -rac
 commands:
                                0/UNRESTRICTED, 1/RESTRICTED
    -pl   --power-limit=        Specifies maximum power management limit in watt
s.
    -cc   --cuda-clocks=        Overrides or restores default CUDA clocks.
                                In override mode, GPU clocks higher frequencies
when running CUDA applications.
                                Only on supported devices starting from the Volt
a series.
                                Requires administrator privileges.
                                0/RESTORE_DEFAULT, 1/OVERRIDE
    -am   --accounting-mode=    Enable or disable Accounting Mode: 0/DISABLED, 1
/ENABLED
    -caa  --clear-accounted-apps
                                Clears all the accounted PIDs in the buffer.
          --auto-boost-default= Set the default auto boost policy to 0/DISABLED
                                or 1/ENABLED, enforcing the change only after th
e
                                last boost client has exited.
          --auto-boost-permission=
                                Allow non-admin/root control over auto boost mod
e:
                                0/UNRESTRICTED, 1/RESTRICTED
   [plus optional]

    -i,   --id=                 Target a specific GPU.

  UNIT MODIFICATION OPTIONS:

    -t,   --toggle-led=         Set Unit LED state: 0/GREEN, 1/AMBER

   [plus optional]

    -i,   --id=                 Target a specific Unit.

  SHOW DTD OPTIONS:

          --dtd                 Print device DTD and exit.

     [plus optional]

    -f,   --filename=           Log to a specified file, rather than to stdout.
    -u,   --unit                Show unit, rather than device, DTD.

    --debug=                    Log encrypted debug information to a specified f
ile.

 Device Monitoring:
    dmon                        Displays device stats in scrolling format.
                                "nvidia-smi dmon -h" for more information.

    daemon                      Runs in background and monitor devices as a daem
on process.
                                This is an experimental feature. Not supported o
n Windows baremetal
                                "nvidia-smi daemon -h" for more information.

    replay                      Used to replay/extract the persistent stats gene
rated by daemon.
                                This is an experimental feature.
                                "nvidia-smi replay -h" for more information.

 Process Monitoring:
    pmon                        Displays process stats in scrolling format.
                                "nvidia-smi pmon -h" for more information.

 NVLINK:
    nvlink                      Displays device nvlink information. "nvidia-smi
nvlink -h" for more information.

 CLOCKS:
    clocks                      Control and query clock information. "nvidia-smi
 clocks -h" for more information.

 ENCODER SESSIONS:
    encodersessions             Displays device encoder sessions information. "n
vidia-smi encodersessions -h" for more information.

 FBC SESSIONS:
    fbcsessions                 Displays device FBC sessions information. "nvidi
a-smi fbcsessions -h" for more information.

Please see the nvidia-smi documentation for more detailed information.

C:\Program Files\NVIDIA Corporation\NVSMI>

 

Getting ready to head to the airport in a bit. Thought I would upload this quick and sloppy video showing power draw with stock clocks on the GPU using the Galax 2000W vBIOS.

 

If you have the means to do so I'd say go for it... that card is more than adequate for a solid build.

I'm well aware of the convert file and my drive is also on ECFM12.1.

My case has plenty of air flow lol. Airflow is not the issue it's the drive... 14x Fans so there's no issue there. I've already mentioned that it has really good air flow in my earlier replies.

Regardless, the fact of the matter is that this 2TB drive is no match for professional work as I've shown in the extensive tests that I've already provide for you. The reads are great, but the writes are horrible and the 2TB sabrent does get hot.

I've quickly tested the 256GB sabrent for giggles and here are the results. Again, reads are great, but the writes are terrible and very inconsistent in general. This is the same that we saw with the 2TB sabrent drives... so to say that these cheap drives are anywhere near Samsung is not realistic.



~300 MB/s writes is comedic for an "NVMe" drive regardless.. that's slower than avg. SATA speeds. That is more than enough to call it "broken," because it is due to the inconsistency and laughable write speeds. @jaybee83 Interesting eh? LOL .. .My old PM961's from 2016 have faster writes than these.

I'm running the 2TB as the game drive at the moment, but I've already reach out to the company for an RMA for what it's worth, not that I think it'll fix anything, but just for the sake of doing it. Regardless, there are other users who all report similar issues with inconsistent writes with the drive running hot.

The tests that I've done with 32GB write tests x5 runs show it to be slow in the ~1000 MB/s range, which is why it seemed off that you were showing speeds of 2.3 GB/s on a 170GB file, but like @jaybee83 and I have already mentioned there is a difference between copy/pasting, location and with the type of write sequence, let alone that the 170GB was a perfect simulated file with perfect data sequences.



Anyhow, again to each his own. If you're happy great. That's all that matters...maybe the 1TB sabrent is the sweet spot for this particular drive, but even from your initial post of showing the drop in write speeds proves that these are mid tier drives at best. Nothing more.

To be clear I really like sabrent products for what they are. I have several of their trinkets that I use often like the external NVMe m.2 enclosure to SATA adapters etc... well and their NVMe SSD's that I'm currently using.

My point in reiterating the facts here is to make sure that people aren't getting misled by thinking these mid tier drives are somehow on par with the upper tier drives, because they are not. This matters more so to the professional users of course, but it's a less of a factor for the average user as any SSD would suit them well in reality...

At this point this beating a dead horse, but like I've said if you're happy then awesome. Enjoy our SSD.

I agree... if users want security, Hello Linux, goodbye Window$.

It is pretty discouraging indeed. Once you hit a ceiling with the current hardware, it can lose it's fun factor quickly, which is why with the newer crippled hardware, you almost have to catch and release to keep the OC excitement flowing... Looking forward to what the 2080Ti Super cards bring to the table...

For most enthusiasts a set up that is impractical is normally the natural environment, because the set up is used for benching. LN2 and DICE is arguably impractical indeed, but those are never meant to be a daily gaming rig in the first place.

Ex. I can't justify ever using my Dark boards for a daily rig as those are seen as my track cars... IMO its way too overboard to use them as a daily gaming system... it's just too good. Same goes for any CPU over my 9900K's ... it'd be just impractical to use Skylake X chips for a rig just for gaming and web browsing.... therefore, it can also be impractical from that end too...

Now this doesn't mean that you can't use high end goodies to game, but if a build ever gets to the point of being totally impractical then now it's in a whole different category.

I've ran into this situation with my regular gaming desktop that's on water, where I did all my benching on so far...which wasn't even on a test bench. I'm having to keep myself from turning it into something that it was not meant to be lol... The goal initially was to showcase results of what it could do in pure desktop form, but unfortunately at the end of the day it's the numbers that matter so it makes all the sense to go all out.

I'm now reworking my gaming rig so that I can enjoy it for what it is and leaving everything else to the test benches...this includes maybe even putting my 1080Ti in there and calling it a day as I don't need much to run the games that I do play. Heck...maybe I'll just put in one of my 8086K's or 8700K's in there too ... it's way more than enough.

It sure is an expensive hobby indeed and less desiring when things are getting more and more borked in ratio to the money being spent. Sad reality really...

I've been thinking about this a lot lately too. What sounds even more enticing is pushing mid range hardware over the latest and greatest as it'd be more cost effective and in someways more rewarding...

I second that... @Vasudev desktop time.

Totally agree brother... we're in a plateau period while we watch AMD dance in the rain of glory for the time being until Intel drops the mic with their new offerings... either way, I'm glad to see Intel not alone anymore in that sector... competition is good.

I'm really curious how big of a bang Intel will make with their debut in the GPU market as well...

@Mr. Fox perhaps stepping away and getting your mind off of things will help reset everything like what @Rage Set said... wait it out to see how everything unfolds with Intel's new goodies...

Excellent board... you'll enjoy it a lot. It's very strong.

It does sound like a carefully crafted blanket statement to satisfy the average consumers. Can't blame the guy..more power to him because that's a bait that 99% of the consumers will bite on, ultimately resulting into buyer satisfaction whether it is true or not...

There are many good things going for AMD and I'm happy to see them back on the map, but aside to the bells and whistles with their new line up, the story is still the same in that AMD is a lot of bang for the buck.

For the average gamer, which is pretty much the majority, AMD purely on a cost stand point is the better choice regardless of the lower clock speeds. However, with Intel lowering their prices, it changes up the playing field quite a bit in that for a large percentage of the gamer crowd, the desire in having the ability to overclock is all a part of the fun factor, which AMD clearly lacks...

IPC, pcie 4.0 etc... on the newer boards, great. But...does that really matter right now for the average person who just want's to game? Likely not...

So the way that I see it is the flashy bells and whistles of what AMD is offering is like NVIDIA with their RTX.... neat, but not a necessity by any means, which makes AMD in reality right where they've always been... budget focused and a lot of bang for the buck.

Fair point, but if AMD is really indeed maxing out the chips then the low price point justifies its limited clock speeds. This isn't a bad thing at all.. it's just one of those things where it is what it is...

As for the 12 core 3900x vs the 8 core 9900K, when looking at it from a price stand point sure... but from a hardware perspective of course more ponies will fair well with multi-tasking... When it comes to over clocking, so far the 9900K with 4 less cores is surely holding it's own in showing what is possible.

The point here is that AMD's scalability is poor compared to that of Intel.





This doesn't meant hat AMD is bad, rather it just shows where they are lacking. Now whether this supports the guys statement of " it doesn't overclock well because "we've already fully optimized it so you don't have to" or not, the fact remains that AMD does not have the fun factor of being able to OC like what you can with Intel. This also does not mean that there is wasted potential left on the table with Intel...

Now as we see Intel lower their prices, having the extra head room closes the price to performance gap quite a bit....

It's all relative... interesting stuff nevertheless...

Note: I started off as Team Red way back in the days and have an AMD build with currently the 1800x in the guest room so I role with both Red and Blue... It's just that Team Blue dominates in the areas that interest me. Could we see that change? We'll have to wait and see...

Very interesting!

Speaking of sound cards...I've been curious if anyone has tried EVGA's NU Audio Card?

Great stuff thanks for uploading this. Have a great trip and hope you get some good R&R.

 

https://www.3dmark.com/3dm/37669381

https://www.3dmark.com/3dm/37643141

Decent results.

 

Well, about your 11 case fans & ensuring enough airflow, more fans doesn't necessarily mean more airflow, it can actually make things worse, I tried different case fan configurations for my case with anything from 2 to 6 case fans, and 3 total case fans provided the lowest temperatures of the different components, so more fans does not mean better temperatures or better airflow over your components (incl the NVMe SSD that we're talking about). Chances are the 2TB drives heats up more than the 1TB drive though, because like I said temperatures are fine with my 1TB Sabrent Rocket NVMe drive (maximum of 60 degC). And it's true that the 1TB version is indeed the sweet spot for this product, like I said before re performance - it has the highest performance out of all the different sized drives. I think it was a bit disingenuous to hammer the 256GB Sabrent Rocket with negative press, when we know it's obviously gonna be the slowest of all the drive sizes when it comes to writes, so I find that a bit disingenuous. They're perfectly good drives for the majority of users out there, bar some of the professional users you talk about, so describing them as "broken" is wrong, and again disingenuous. In some things these drives are faster than the Samsung Evo's even, it depends what you're doing, they're certainly not broken.

 

@iunlock Its time for Ryzen 37xx or 39xx or even 3rd gen Threadrippers. I want it to be a good purchase with Air/CLC cooling and this time I'll see if 40x or 45x is achievable or not. I'm tired of choosing a perfect laptop which there isn't any!

 

... much elided ...

If you're talking about overclocking as an end unto itself rather than a means to achieve more productivity/frame rate, what have you, then that's a different matter (but note that someone apparently achieved 5.2+ GHz all core on a pre-production 3950X... with liquid nitrogen cooling). But that's certainly not what I think of -- I think of it as a way to achieve better performance with the same outlay.

It sounds like you'd be happier if AMD sandbagged their base and boost clocks and intentionally left something for people who want to be really clever to achieve. To me that's perverse; I'd rather get the best possible performance out of the can. What I'm seeing reported is that on MT workloads even the 3700X is mostly beating the 9900K by a good margin and the 3900X is blowing it out of the water. On single threaded workloads it's a lot closer, and gaming (which I don't care about personally) does tend to favor the 9900K, with a much smaller margin than with the 2700X. But I read (and can't find right now) that that may be due to memory latency, which faster CPU clock rate isn't likely to fix.

As for comparison against the 7960X, that's priced almost 3x the 3900X, has 1/3 more cores, and (for what it's worth; I understand that that's a somewhat iffy measurement) much higher TDP. I'd hope it would be quite a lot faster, but it isn't.

Clock rate is far from the most important factor where it comes to performance.

 

It's actually 14x case fans, but the point is that it seems like your doubting my knowledge of the simple dynamics of air flow? I along with many here have been in this game long enough to understand the proper orientation for good and sensible air flow depending on the set up...static pressure etc...

This is about the 4th time again that you've referenced air flow and using that as a possible cause of the high temps, when in reality it has nothing to do with it.

The point is... the Evo Plus was tested under the same conditions regardless if there is good air flow or not... so the testing environment is the same and consistent. (1:1)

The point about the 256GB is that the write speeds are below sub par, because it is. I've referenced how it's worse than my 2016 PM961's that I use in some systems. For a 2019 m.2 NVMe drive, regardless of the storage capacity, the facts are very clear... the sabrent drives are mid tier at best.

Using the word "broken," is common to describe not only something that doesn't work, but also something that functions poorly or way below par to where it should be.

I don't know about you, but ~300MB/s writes or even ~1000MB/s writes on a drive that's a m.2 NVMe is not impressive and IMO not worthy to wear the badge to even be called an NVMe m.2 drive. Again, that's just my opinion...

I've mentioned numerous times that the sabrent drives are more than adequate for most people lol... I feel like a parrot repeating myself over and over again.

Let's not forget that I own the sabrent drives as well and use them daily... the 2TB sabrent is my main gaming drive so it's not like I don't own them to not know what I'm talking about.

I've also provided extensive tests and data to show the simple facts regarding these drives compared to the Evo Plus, in response to your initial curiosity / post.

As you and others have seen the writes of the sabrent drives are not up to par, putting them exactly where they belong...mid tier drives at best.

There is nothing disingenuous about any of this, rather quite the contrary...

I'm basing everything on facts, real world experience and data; not what I want the drive to be or based on feelings. That has nothing to do with any of this. It's just good ol' data that we can all see for ourselves to make a sound decision based on the usage and needs.

There's the famous saying "It is what it is..."

I'm not stuck to any brand as I own pretty much all of them.

I like what I'm seeing with the new pcie 4.0 drives from sabrent, corsair and aorus so far (all the same under the hood as we know) ... there's still a quite a bit of head room with pcie 4.0 so I'm curious what Intel, Samsung or others will do with their pcie 4.0 drives.

Like I've said before... if you're happy then awesome. Stay happy, enjoy your drive and carry on...

Great choice... I'm set on the 3900X ... before I was going to just go with the 3800X, but since I already have plenty of 8 cores ...going 12 cores would be more refreshing. It's crazy how it's all sold out here in the States.... but as expected. Good for AMD... I'm glad they are eating their cake right now, because it forces Intel to answer back... but we won't be seeing that anytime soon with anything that is earth shattering.

For the price I think the 3900X is the sweet spot in the line up... you know how to reach me if you have any questions. I've been doing a lot of homework for this Ryzen build so I'd be happy to share...

Precisely and that's what I have been making very clear in distinguishing the difference pure benching vs real world usage / tasks.

There's no question that going LN2 is impractical for real world stuff... but that's very obvious... LN2 is very practical for the benching world for that class, but that's a whole different thing all together.

As for AMD and if I'd be happier if they sandbagged their base and boost clocks intentionally.... that's incorrect and not true.

I'm aware of AMD's approach vs Intel's .... they have very different play books between the two.

AMD makes all the sense for the majority of consumers and it's a no brainer (IMO) to go with the 3900X for example at $500 vs the 9900K $500 .... however, as we know with the price drops, it's closing the gap in the areas that normally gives AMD a huge lead when it comes to performance per dollar. They still have an edge, but it depends on the user and what they are after when choosing between the two.

Edit: The Ryzen Threadripper 2950X (16 Core / 32 Thread) 4.4GHz Max Boost - Is currently on sale for $499 so in that price range, wow ....

As for the 3700X vs the 9900K ... there are tests (9900K vs 3900X 4 cores turned off) done by credible sources that clocked them both to 4.0GHz to test the ipc etc... The 9900K actually still comes up on top overall. So it really depends on the scenario ... for gaming, I'd still go with the 9900K over the 3700X or 3800X any day.

When it comes to 'price to performance' of the upper tier chips, rather for the chips at 12 cores and above... AMD has advantages due to the price alone... but it's important to keep in mind that a lot of tests being done are stock for stock ....

Regarding that link... 7960X? That was stock clocks lol... Let's have Brother @Mr. Fox who has the 7960X blow that score out of the water and show the REAL scores and capability. I'm sure he'd be more than happy to...

Remember...Intel can OC and scale a lot better in general, whereas AMD can not.

So stock for stock it might be close... but it all changes when Intel boosts their clocks on ALL cores.................... because they simply can.

For the majority of us enthusiasts we consider a true OC when it is on ALL Cores.... compared to just one. (Again...generally speaking.)

Anyhow... fun stuff... I'm just glad that we can be having conversations between AMD vs Intel lol... it's about darn time...

 

Well his build is working well thus far. We decided not to upgrade CPU for now. The AMD R9 270 is holding up at 1080p/720p resolutions on his TV. Next step will be overclocking the i7 950 to 4Ghz. I forgot how much harder it was to OC X58 LOL. I did get it to 4.2Ghz a long time ago.

He's getting a new toy this weekend. MSI GT73VR with i7 6820HK & GTX 1070. Will be fun to play with. Hope that CPU can beat his OC 3820QM LOL.

 

Well, I think I've made some valid points, and I think you're over egging it in general, but we've both said what we want to about it.

 

A little off topic but being a Marine I really enjoyed this. Maybe you will too.

 

Very nice... The prices for CPU's right now are really good.

Ex. Ryzen 7 2700X for ~$200 and even the Ryzen Threadripper 2950X 16-Core for $499 ...

Keep us udpated with the GT73VR. Looking forward to it.

Over egging? How so, by providing facts? LOL

Last I've checked you asked... I provided answers with detailed tests... then you go on a tangent to try and make something that is mid tier to seem better than it is?

I think you're taking it too personal which is silly. But anyway... facts beat feelings.. it is what it is...

There's nothing really more to talk about regarding this subject. I've already laid out all the facts.

Enjoy your SSD.

 

Hey, I think I've said everything I want to already on this, so you're just gonna have to lump it. ;-) Thanks for doing the initial testing though.

 

You're welcome.

 

good forum just buy this monitor:
https://www.amazon.es/gp/product/B075B1XCBY/ref=ppx_yo_dt_b_asin_title_o00_s00?ie=UTF8&psc=1

and to my amazement I have it at 2560x1440 144hz .... experts who say if you can **** something on the monitor to have it with oc ... xd ... but at 2k it looks fable, I think the same 2k is worth 500 euros

 

https://www.3dmark.com/3dm/37820584?
Best I'm able to accomplish on my current desktop it seems, cant break that 34K Barrier.... Any Suggestions?

 

Have a look at the following link, it's showing a decrease in performance of AMD 3xxx CPUs when undervolted even if clockspeeds are remaining at the same Ghz figure as reported in monitoring software:
https://www.notebookcheck.net/Under...ng-36-less-power-at-1-00V-Vcore.427867.0.html

Strange!

I had wondered if undervolting AMD's new CPUs would actually increase performance due to greater thermal headroom allowing for greater 'auto boosting', but it looks like you just end up with less performance instead (well ok, and less power consumption too).

 

Test with different graphics drivers, Use InSpectre tool to disable security patches, lowest possible GPU temp. Maybe tweak and or increase memory speed, as well look after if there is something that eat cpu clock cycles (Disable unnecessary Services, Software, Bloatware, Anti Virus). Test with other Futurmark versions etc. A lot you can try.

https://www.3dmark.com/compare/fs/19917477/fs/19521970#

 

I also compared with Premas P870... Maybe the 9900k is throttling? Stays under 70C at least from what I can see on the OLED Panel on the board, maybe I'll run a test with HWINFO Open, Currently using XG XTREME Drivers 431 with pretty much everything removed except 3 packages HD Audio NVI2 and the display driver. Running Windows 10 Lite and have further disabled and removed services. Memory speed I tried tweaking its not really booting well to an OC its 16GB Per DIMM and IK it doesn't like high clock speed as much as 8GB Per DIMM does.

Sent from my OnePlus 7 Pro using Tapatalk

 

Try run with single card (disable Sli)and run same Cpu clock speed. If you get same Physics score with that high cpu clock speed something is clearly wrong. Test also with stock Cpu. Other graphics drivers will as well make a difference.

 

Running full custom loop would have to drain loop to remove one card so would rather not do that..

Sent from my OnePlus 7 Pro using Tapatalk

 

Can't disable SLI other way? https://www.computerhope.com/issues/ch001167.htm

 

Through the driver or removing the bridge...

Sent from my OnePlus 7 Pro using Tapatalk

 

The main point is... See if you get proper Cpu score vs. Clock speed used. As well compare your graphics scores vs reviews from same card (default + same overclock).

 

When comparing my score currently to setups around it at the same score looks like cards clocked much lower than mine, I have a 1000W PSU could it be I'm out of power to feed the cpu and the 2 cards under Load so it's throttling?

Sent from my OnePlus 7 Pro using Tapatalk

 

1000w psu with 1080Ti Sli without volt mod should be enough. But remember the Cpu run alone in the Physics test. Run max Gpu overclock with stock Cpu. Kill-a-watt to measure max power from the wall is also something you could test with.

 

Stock core clock +800 mem gives a score of 32.3K

Sent from my OnePlus 7 Pro using Tapatalk

 

I talk about the Physics score (Cpu score).

 

Stock GPU core with +800 MEM and CPU set to Auto gives 32.4K Overall and 24.6K Physics.
Overclocked GPU and CPU set to Auto gives 32.9K Overall and the same 24.6K Physics.
Both Overclocked Overall 33.6K and 26K Physics.

Sent from my OnePlus 7 Pro using Tapatalk

 

CPU Max temp in Auto (5GHz 1.35V) Max Temp of 86C of the Hottest Core Max and Min Clock Speed of 5000MHz...

Sent from my OnePlus 7 Pro using Tapatalk

 

You shouldn't get same Physics score with stock and oc'd processor. Not sure if you have the paid version. If you have you could run Custom setup with only physics test (test Cpu stock and oc'd).

 

I'm not when stock 5GHz I'm getting around 24K with 5.2GHz 26K

Sent from my OnePlus 7 Pro using Tapatalk

 

Stock 9900K is 4.7GHz on all 8 cores... https://www.3dmark.com/fs/19385732

 

My boards Auto sets it to 5 Regardless it still goes up in Points as the clock speed increases