Jump to content

??? LINUS OPTANE CONFUSION ??? House of Quality? Customer requirements? Google? Apple? BIT COIN??????

Dr. KEK

but the whole optane thing is still about IOPS. Why the hell do you need throughput to grab 400,000 separate .asm files less than 1kb in size? I need IOPS to shove into RAM, compile, and dump the exe back to disk. 

 

I can post more in depth later about how I have overclocked the Uncore on my Xeon to increase the IOPS on a custom PCIE 1x NVME adapter I made myself. Intel uses 2x, so no overclocking needed. 

 

as for the average user workload, and who Optane is marketed to... you have CLEARLY NEVER LOADED A COMPLEX SOLIDWORKS ASSEMBLY. There are THOUSANDS of small .SLDPRT files to grab from disk. Endless workstation loads need IOPS from optane and not necessarily raw capacity or throughput.

Link to comment
Share on other sites

Link to post
Share on other sites

*Thread moved to the General Discussion*

If you need help with your forum account, please use the Forum Support form !

Link to comment
Share on other sites

Link to post
Share on other sites

LTT is not the forum for technical discussions of devices like this. You want to get into the dirty details, join ArsTechnica. They have actual Intel engineers posting there. LTT is for fanboys of OCing and wrecking expensive parts.

NOTE: I no longer frequent this site. If you really need help, PM/DM me and my e.mail will alert me. 

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, huilun02 said:

But I can't think of anyone on this planet who doesn't use Solidworks

Silly Linus, forgetting the most important software that his professional team uses everyday

what's solidworks?

"If a Lobster is a fish because it moves by jumping, then a kangaroo is a bird" - Admiral Paulo de Castro Moreira da Silva

"There is nothing more difficult than fixing something that isn't all the way broken yet." - Author Unknown

Spoiler

Intel Core i7-3960X @ 4.6 GHz - Asus P9X79WS/IPMI - 12GB DDR3-1600 quad-channel - EVGA GTX 1080ti SC - Fractal Design Define R5 - 500GB Crucial MX200 - NH-D15 - Logitech G710+ - Mionix Naos 7000 - Sennheiser PC350 w/Topping VX-1

Link to comment
Share on other sites

Link to post
Share on other sites

Just now, SC2Mitch said:

really? i never knew a linus tech tips forum would contain such fans, i'm deeply shocked by this! (i really hate linus!)

of OCing...apparently you missed that part...

NOTE: I no longer frequent this site. If you really need help, PM/DM me and my e.mail will alert me. 

Link to comment
Share on other sites

Link to post
Share on other sites

1 minute ago, wkdpaul said:

it's a CAD/CAE software.

I know... I just saw the opportunity and took it xD:P

"If a Lobster is a fish because it moves by jumping, then a kangaroo is a bird" - Admiral Paulo de Castro Moreira da Silva

"There is nothing more difficult than fixing something that isn't all the way broken yet." - Author Unknown

Spoiler

Intel Core i7-3960X @ 4.6 GHz - Asus P9X79WS/IPMI - 12GB DDR3-1600 quad-channel - EVGA GTX 1080ti SC - Fractal Design Define R5 - 500GB Crucial MX200 - NH-D15 - Logitech G710+ - Mionix Naos 7000 - Sennheiser PC350 w/Topping VX-1

Link to comment
Share on other sites

Link to post
Share on other sites

update: we ordered an expensive Xeon workstation to compare, with a real Quadro card inside.

 

The 2-socket 8-core (2x 4core chips, with HT, and quadro gpu) runs a bit slower than my i7 7800x setup(and rx580) according to the official solidworks performance test.

 

 

As for this X58 setup, it is comparatively on par, surprisingly close to the Xeon workstation we ordered. It was manufactured in February 2018 :P

Link to comment
Share on other sites

Link to post
Share on other sites

Can't tell if troll or...

 

We've never said Optane technology makes no sense.

 

Intel is marketing this drive as a boot drive, which doesn't make a ton of sense compared to just buying a bigger SSD and a smaller/cheaper Optane module (which has already existed for nearly a year).

Link to comment
Share on other sites

Link to post
Share on other sites

22 minutes ago, LinusTech said:

Can't tell if troll or...

 

We've never said Optane technology makes no sense.

 

Intel is marketing this drive as a boot drive, which doesn't make a ton of sense compared to just buying a bigger SSD and a smaller/cheaper Optane module (which has already existed for nearly a year).

I can't see rebooting times being a critical factor in decision making any longer. Any SSD is fast enough for booting a workstation. Maybe I misunderstood their marketing altogether. Our next build will be a Mini-ITX i7-7820x and quadro. Optane or other NVME as boot, as well as the workplace applications we use(80-120gb in program files. user docs are on server). The boot drive contains the programs. It's a ghost image, basically... a domain workstation. It loads programs from nvme and works on files over 10gbe.

Link to comment
Share on other sites

Link to post
Share on other sites

14 minutes ago, LinusTech said:

Can't tell if troll or...

 

We've never said Optane technology makes no sense.

 

Intel is marketing this drive as a boot drive, which doesn't make a ton of sense compared to just buying a bigger SSD and a smaller/cheaper Optane module (which has already existed for nearly a year).

home crap computer vs build at work lol

xeon_x58_performance.png

Link to comment
Share on other sites

Link to post
Share on other sites

according to my calclations the MIPS ratio is 1.5804878048780487804878048780488:1, favoring the 7800x to westmere-ep. so multiply the 7800x times by 1.5804878048780487804878048780488 and shoudl match mine for the most part.

 

PROCESSOR EXPECTED: 41.88. actual: 38.9

I/O expected: 28.75. actual: 30.8

overall expected: 91.8. actual: 101.1(my GPU is worse at home, RX480...)

 

The gap in simulation I think would have to do with hyperthreading not being enabled at this time, as I have switched down from 4.5GHz HT ON(winter clocks) to summer clocks, 4.1GHz HT OFF.

 

The previous measurements were single-threaded heavy weighted. The simulation is purely multithreaded performance, which is why I mentioned hyperthreading in the first place.

 

So what did we learn?

 

-2018 IS HERE AND SINGLE THREADED PERFORMANCE IS STILL KING

-PLS CAUTION AS MULTITHREADED BECOMES MORE SNEAKILY PREVALENT 

-lol AMD

 

 

Link to comment
Share on other sites

Link to post
Share on other sites

Hey people need to shave off 24.8 seconds in order to eat their Double Big Macs and XLarge Diet Coca Colas and XLarge fries. While they work in Solidwerx and surf porn and surf LTT

Link to comment
Share on other sites

Link to post
Share on other sites

7 hours ago, Canada EH said:

Hey people need to shave off 24.8 seconds in order to eat their Double Big Macs and XLarge Diet Coca Colas and XLarge fries. While they work in Solidwerx and surf porn and surf LTT

A typical simulation can run 6 minute to hours, depending on the granularity of the data you need. The time span reported is the solidworks benchmark time, just like Unigen Heaven or any other timed benchmark, where scores are compared by percent difference, and extrapolated for larger projects - such as predicting how long it would take to encode a 4k movie 60min length on one box vs. another. 

 

The numbers we compute are used to design parts of medical devices - minimizing size, cost, weight, etc.  - instead of being used to render your favorite kitten with 8xMSAA for YouTube.

Link to comment
Share on other sites

Link to post
Share on other sites

8 hours ago, LinusTech said:

Can't tell if troll or...

 

 

I can. And it is obvious.

Link to comment
Share on other sites

Link to post
Share on other sites

3 hours ago, RollinLower said:

i like the super clickbait of the title tough. 11/10 top memes.

I was trying to emulate the new YouTube Thumbnail mentality, only in Forum Topic form. I thought it would align with the direction their organization is headed.

Link to comment
Share on other sites

Link to post
Share on other sites

4 hours ago, Dr. KEK said:

I was trying to emulate the new YouTube Thumbnail mentality, only in Forum Topic form. I thought it would align with the direction their organization is headed.

Well, it worked. I clicked on your thread :P

Link to comment
Share on other sites

Link to post
Share on other sites

update: i was experimenting with my NB speed and QPI speed. 

 

My motherboard and CPU is way too old to run extended QPI speeds reliably, I tried juicing all the voltages with little luck for stability. Specifically I boosted IOH voltages from 1.1v all the way up to 1.5v incrementally.

The motherboard has been powered on since 2009. 

For testing, it is reliable, so I got some benchmarks worth noting...

 

I increased NB to ~3.95GHz, CPU was ~4.10GHz, I tightened the RAM down to 7-8-7-20 1T.

This boosted my 1x NVME score past 17,000 IOPS in HD Tune random access.

The 4x NVME score went up accordingly. 

 

I am wondering if the difference in IOPS is due to physical location of my NVME cards. The 1x slot is about 3cm from the actual Southbridge, 5cm from the CPU(IOH ON CHIP).

 

The 4x nvme is plugged in the 2nd PCIE x16. It is over 10cm from the CPU, 5cm from the ICH/southbridge.

 

I don't have PCIE 2.0 validation equipment any longer, it has been years. 

Link to comment
Share on other sites

Link to post
Share on other sites

image.png.7fcdf65f65eb7f3fb1f1522eb6a000b3.png

 

 

 

This seems to be my resting summer clocks until x299 proves it is around more than a year or two. I have hyperthreading off, as f*$@ing Rainbow Six Siege can't handle it lol

Link to comment
Share on other sites

Link to post
Share on other sites

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 account

Sign in

Already have an account? Sign in here.

Sign In Now

×