>Went from an i7 5820k at 4.7GHz >To a Ryzen 1700x at 3.9GHz

>>61086187
Stuck with 2500k, I'm WAITING™ for Zen2

>>61086187
Nice try amd

File: ThreeDimensional 560.png (405KB, 358x500px) Image search: [Google]

405KB, 358x500px

>>61086187
is that a boy

>>61086308
https://www.youtube.com/watch?v=2xRAACegBzE

File: 1479192198604.jpg (324KB, 882x758px) Image search: [Google]

324KB, 882x758px

>>61086187
remoov

I bet you broke your shit and were too poor to get an upgrade so you settled for AMD.

>>61086308
Oh dude, Zen2 will be a midget upgrade form Zen1. WAIT™ for Zen5 in 2022.

File: dumb tripfag gets btfo.png (2MB, 1312x1740px) Image search: [Google]

2MB, 1312x1740px

>>61086457
When will you kys yourself?
>>61086458
Zen 2 comes in 7nm with 5 Ghz operation in mid 2018

File: blocked and reported.png (16KB, 741x185px) Image search: [Google]

16KB, 741x185px

>>61086457
did someone steal his trip or was this retard always so obnoxious?

>>61086509
I filtered him earlier.

Literally nothing but garbage comes from his mouth

>>61086509
Since it's clear how much better Zen is compared to Intel alternatives, the shill has become belligerent. No more talking points, just kicking and screaming.

>>61086505
>in mid 2018
You're dreaming

>>61086418
- wearing a wig
- highly visible philtrum
- the fingers are quite long and not tapered like a girl's

based on these visual cues, this is 100% girl (male)

>>61086582
Why would you buy a 1700X when you already had an >i7 5820k at 4.7GHz?

That's a rhetorical question, you wouldn't.

>>61086745
I did it because i wanted something new, and i want to support AMD. Other than that, AM4 is not a dead platform, unlike x99

>>61086745
Why would you still post here when you're an unintelligent baboon that gets called out on the stupid shit you say on a daily basis?


That's a rhetorical question, you wouldn't.

>>61086754
Because its fun.

>>61086767
I know you hear this every day but filtered.

>>61086745
>>61086767

Oh, and im selling the i7 stuff for almost 750 USD so im cool with it lol

File: strawman.jpg (26KB, 400x400px) Image search: [Google]

26KB, 400x400px

>>61086754
Also.

>>61086775
The fact that you feel the need to announce it means I already won.

>>61086187
>multiple other intel/amd posts
Can you go shit up one of those instead of taking up more space?

File: 1480901923125.jpg (328KB, 810x587px) Image search: [Google]

328KB, 810x587px

>>61086582
True

>>61086708
Women have visible philtrums too you assfuck

File: uwotm8.png (152KB, 2440x595px) Image search: [Google]

152KB, 2440x595px

>>61086803
Not him but I'm not gonna filter you. I recognize you as the official clown of /g/. Shame others are too autistic to bask in the irony of your posts.

>>61086862
Is him the dumbest tripfag ever in /g/?
I mean, even InElite understood basic stuff like this

>>61086862
Says the guy that thinks an AMD and Intel laptop watt hours are equivalent.

Fucking retard lol. Its like if you don't get a response you get stuck in some autistic loop.

>>61086888
He's just pretending. He gets a rise out of it and the ones who know his ruse as well. If he was actually that retarded he would have quit posting here years ago.

>>61086906
lmao, I know right? I don't think downing a bottle of riddlein can fix me. Hell, 10 lines of coke wouldn't even touch the tip of the iceberg.

>>61086187
Nice. Having better minimums means nearly everything for games. Smooth framerate > roller coaster of choppy trash. I'd take a smooth 30 over unreliable 60 any day.

File: 1489510914590.png (421KB, 475x798px) Image search: [Google]

421KB, 475x798px

How could have intel not seen zen coming and why did they spend the last 5 years jerking each other off in the board room instead of working on a significant arch after sandy bridge?

>>61086937
I really don't think you should be taking stimulants if you really are suffering the way I think you may be.

>>61086972
They just got lazy as hell and used to ripping off their customers. That is the world should AMD ever drop out of the CPU market. Fuck that and fuck all that fanboy shit DESU. Both companies are important and the more competition between Intel and AMD the more we win. Better products, better prices etc. Tards who trash on AMD are content with getting taken advantage of and shit on by Intel. No thanks.

>>61087003
If you were a stimulant, I'd take you desu :^). I'd probably die of laughter while jerking it on my balcony watching as people get covered in my cum scream in horror.

>>61086912
I never lie or put on false personas, everything I say and do I mean whole heartedly at the time. I basically don't give a shit and have never understood attention seeking or validation seeking type behavior.

I post about shit I'm interested in and I like to berate people constructively. If it comes off as inflammatory its probably only because I've said something that you took personally because you are a tool.

>>61086187
> changing CPU to get 5% perfomance increase
Congratulations on contracting consumer disease.

File: old.jpg (35KB, 420x287px) Image search: [Google]

35KB, 420x287px

>>61087068
10/10 pasta

>>61087068
This is a nice pasta my senpai, the only quality post you ever made

>>61087080
>implying the ryzen 1700 doesn't absolutely smoke the 5820k in multithreaded performance

>>61087080
Nah, he got like 25% better multi-threading. Better than the skylake to kaby lake jump intelcucks did.

>>61087068
You're saying you're not pretending to be an idiot, so you really are one

>>61087092
>>61087087
Only sub-human retards are dishonest. Its a social hack for people that can't cut it that erodes their sense of self worth until there is nothing left. Their only comfort is their own denial that honest people actually exist.

>>61087125
I can be an idiotic sometimes but I wouldn't consider myself an idiot.

>>61087125
Or maybe he's the simple jack of 4chan, bound to walk this earth for all eternity in a slumber of stupor he laid upon himself in hopes of seeking the ultimate fame and glory.

>>61087139
>>61087191
This is top tier banter my senpai

>>61087139
Preach it brother.

>>61087204
batter*
If it materialized in front of me I would slather it all over my left nut and nut for the sake of nutting.

File: weeez.jpg.jpg (19KB, 222x293px) Image search: [Google]

19KB, 222x293px

>>61086419
>ex commie

No thanks

File: f7f7ef742b89401f70f92f02b4f25128.jpg (50KB, 700x900px) Image search: [Google]

50KB, 700x900px

>>61087191

Im still using an fx 8350

>>61087298
Nice, you're the man with a plan. Hate how everyone is upgrading every fucking year because intel or AMD overhyped the latest abomination they rushed out of the basement.

I'm getting an r5 1700 and sticking with it until it dies or a huge significant reason comes along to warrant the upgrade.

>>61087285
Its all a matter of perspective. Conflating spiritualism with religion to me seems a form of idiocy.

Suppose a schizophrenic person believed he was talking to a dead relative in a mirror, would you consider that religion?

>>61087350
There's actually some legitimate reasons to upgrade sometimes.

That includes resell value of old cpu being high and new CPU performance value being high with marginal cost factor (equal to resell value of old cpu or within margin)

>>61087370
Yes, yes I would actually. His firm belief in the mirror man cannot be disproven.

And because I have a mirror in my room.

>>61086308
I'm in the same boat. Was planning to upgrade to Zen, but I ended up buying a new TV.
I'll probably end up waiting for Zen2 now.

>>61087405
I'm just saying you can't automatically dismiss spiritualism as religion because its a far more base idea. You could argue that spiritualism requires religions beliefs but it really doesn't because if it did then religions wouldn't have spiritual beliefs at all, someone had to come up with them on their own.

>>61087448
Spiritualism 99.999999999999% of the time means you worship something and accept something as a higher being which you cannot scientifically verify to be 100% true.

I'm sorry if it hurts your feelings but both things practically mean the same thing for all intents and purposes.

Doesn't mean we should dismiss things like the man in the mirror as fictitious beings since we cannot will things in or out of existence irl but no harm is done when we question their existence altogether.

>>61087448
>he is still posting

File: x265-2.png (582KB, 600x600px) Image search: [Google]

582KB, 600x600px

>>61087631
>userbenchmark
Might as well post pissmark and gookbench scores

>>61087537
>he will continue to until this thread closes

>>61086187
nobody cares about your gaymen life, faggot
go back to /v/

>>61086187
went from i7 4790k
to i7 4790k because you literally don´t need anything else if you are not doing work with it
>inb4 muh gaymes
fuck of with that garbage you can play perfectly fine with a 2500k

>>61087777
My i7 5820k was very capable, but like i said i wanted to try something new, and im getting a good buck out of my old platform so why the fuck not?

>>61087788
because you are wasting money "to try out something new"
Do as you wish if you deem that investment worthwhile but it´s always going to be a waste of money

>durr went from a 6 core 2012 processor to a 2017 8 core processor
>it's better hahah amd is le better xD

>>61087842
q3 2014, you dumb fuck

>>61086187
So you actually wasted $500 to get slightly worse gayming performance? I doubt you're using ""multithreading"" for anything.

>>61088077
>MUH GAEMSS
>>>/v/

>>61088101
Did you even read the OP?

>>61088130
Did you read it? OP gets better minimums and better multi threading, he wins all around

>>61088138
The way he phrased that means he has no idea if he gets better results or not.
>i-it's pretty much the same
Instead of
>I get better minimums and averages

>>61086419
>"in the productivity department even the 1500X will annihilate the 2500K"
>beats 2500K by 1.6%
rly make think

File: 1491615983359.jpg (44KB, 678x381px) Image search: [Google]

44KB, 678x381px

>>61086187
This is the ultimate bang for your buck and all the average gamer requires.

File: axiNi9D.jpg (20KB, 514x488px) Image search: [Google]

20KB, 514x488px

glad op likes his downgrade

>>61088366
2 CORES??? DRRRRRRRRRRRRRUUUUUMMMMPPPFF

>>61086187
Is there any documentation as to hardware backdoors and espionage tools for AMD, like there is in IntLel? I've not used AMD so far, but I am seriously considering it.

>>61088366
>games on a pentium
topkek nice try faggot

>>61086745
Fuck off Jordan, don't you have more product placement to do on logicalincrements.

File: my sides.png (173KB, 640x400px) Image search: [Google]

173KB, 640x400px

>>61088366
>reduced to shilling for the officelady-tier Pentium
Don't you have some spreadsheets to be keeping up to date?

>>61086509
As far as I can hell he's always hated AMD but went full retard when Ryzen was released.

https://rbt.asia/g/search/text/AMD/tripcode/%21BE%2F4wes0mE/
https://rbt.asia/g/search/text/ryzen/tripcode/!BE%2F4wes0mE/

>>61086852
not him, but that is indeed a girl (male)

File: pondering seagul.jpg (38KB, 1024x1024px) Image search: [Google]

38KB, 1024x1024px

>>61088810
>girl
>(male)

>>61086972
Intel didn't expect Zen at all.

AMD isn't a very good company, but Jim Keller managed to make it king. Reminder that he also worked with AMD when it was king 18 years ago.

>>61088835
Keller was project manager on Zen, he left before it was even complete.

>>61088822
feminine penis

>>61088846
He basically completed it all on his own...

If he wasn't there to design Zen, AMD would not have been able to hold a handle to Intel.

>>61088873
haha like you're mum cant handle my penis

>>61086187
she a cute (male)

>>61088835
Zen was made by Mike Clark. Keller was doing K12 design. Both shared some design methodologies.

>>61088368
More of a sidegrade, but that's ok.

File: mpv-shot0003.jpg (67KB, 1280x720px) Image search: [Google]

67KB, 1280x720px

>>61088355
>beats 2500K by 1.6%
That's some fine cherry-picking there, m8ie

>>61089054
at this point I'm just waiting for someone to come in and say "lol it took AMD 8 more threads to get twice the performance of an 2500k in x264 xd"

>>61086906
>Says the guy that thinks an AMD and Intel laptop watt hours are equivalent.
Stupid, right? It's almost like they think conservation of energy is a fundamental law of physics or something.

File: LRM_EXPORT_20170411_145818.jpg (1MB, 1293x1730px) Image search: [Google]

1MB, 1293x1730px

>>61086457
Kill yourself you perma-BTFO'd tripfaggot

>>61086972
I even bought a Ryzen just last month, but you're being stupid. For the most part, Intel is offering better performance. The new i9 7900X may be a housefire, but it's still faster on the vast majority workloads than anything else that has the same 10 cores or less. The real problem with Intel isn't that their hardware is slow, but that they're overcharging so heavily for it. And a bit of housefires, of course.

>>61089192
this needs to be updated with news headlines saying "intel hyperthreading broken" and "i9 7900x pulling 250w under load & 400w when overclocked"

>>61089227
I don't think the extra 25% performance is worth it considering you pay twice as much and that performance is mostly due to pushing the power consumption into housefire territory.

>>61087529
No that's not true. Jordan or whoever the tripfag is is correct in his assessment that spirituality does not require a belief in a higher power or a higher being.

I would also like to point out that your percentage literally means this entire planet. That is anyone who believes in spiritual out on this entire planet.

I also don't understand the rest of your post. It literally makes no sense.

I understand where you come from from a scientific point of you, but you have no Authority I would call it on speaking on any spiritual matters. Do not pretend like you are a master you have nothing. People who pretend to know a lot of things on subjects that they don't really understand don't really have any basis or right in speaking about such subjects.

Thanks for your time. I will no longer continue this.

>>61089261
>I don't think the extra 25% performance is worth it considering you pay twice as much
That's what I said. All I wanted to say is that it's stupid to say that Intel's architecture sucks, when in fact they have the best hardware in terms of absolute performance.

>that performance is mostly due to pushing the power consumption into housefire territory
Too bad you can't even do that on Ryzen, then.

>>61089281
>Too bad you can't even do that on Ryzen, then.
Even attempting to OC on i9 seems futile without exotic cooling solutions and/or delid.

>>61089296
How fortunate, then, that you don't even need to overclock it in order for it to offer the best absolute performance.

>>61089281
>it's stupid to say that Intel's architecture sucks, when in fact they have the best hardware in terms of absolute performance
That's going to change when Threadripper arrives. Really hoping it's sooner than later. Intel needs a good smackdown. You don't play core wars with AMD and win.

What type of people support intel in 2017? I just cannot fathom what happens in their rotten head.

File: amd-zen-core-scaling.jpg (33KB, 597x430px) Image search: [Google]

33KB, 597x430px

>>61089227
>A $999 HEDT CPU is less than 50% faster than consumer R7s
Yes congratulations, however it's not that long until August 10th when Threadripper comes out and utterly dumpsters Skymeme-X.

Remember the 16-core is going to close to 1.99 times an 1800x for $999 or less.

>>61089306
I agree, though of course it also remains to be seen what happens when the HCC i9s arrive later this year.

Again, I'm not at all trying to say that AMD sucks, all I'm saying is that it's stupid to go around saying that Intel's architecture is bad.

>>61089383
Wouldn't say it's bad, more like optimized to the point where further improvements are not likely to happen. They needed their new arch out now, not 3-4 years from now.

>>61089376
>Remember the 16-core is going to close to 1.99 times an 1800x for $999 or less.
You're completely inane if you think that isn't going to depend on the workload. I'm sure there are workloads where that's true, but making any particular statement as to how common they are is just going to make you look stupid later on.

>>61089393
And we don't know how much more improvements AMD can get out of Zen either. It's terribly easy to speculate freely in how much better Zen2 will be, but we don't know this yet.

I'd venture so far as to say that further gains from Zen will be marginal, because both Zen and Core already extract virtually all the intrinsic ILP that is present in serial programs. There's a reason you never run the 8-wide Power8 in less than SMT2 mode.

>>61089395
Yes, practically anything that is multithreading dependent, but that's the majority of HEDT workloads, so it's fine.

>>61089395
Nearly linear scaling in all multi-threaded workloads is literally the sine qua non is the Zen architecture. The 16-core Threadripper is just 2 1800x's dies on one package linked by IF, and due to the huge increase in mass and surface area of the heatspreader there is no reason to not expect 3.6-4.0 clocks on the 16-core.

>>61089406
As long as they can use 7nm and improve IPC by 10% or so, we'll definitely see some improvements worth paying attention to.

>>61089429
>practically anything that is multithreading dependent
You are indeed inane. Multithreading something does not mean that its performance will scale linearly with the amount of processors you throw at it.

>>61089440
>As long as they can use 7nm and improve IPC by 10% or so
Yes, then they'll be more or less on par with Intel. Which is great, but it confirms what I said that the reason Intel can't get much more out of their architecture is because it runs x86 code more-or-less optimal, quite simply, which would then be true for Zen as well.

File: Screenshot_20170626-082956.png (38KB, 1080x1920px) Image search: [Google]

38KB, 1080x1920px

Even as a phone poster, I couldn't handle this much retardation

>>61086745
While it's true it's pretty much a sidegrade, it's wasn't nonsensical to sell the platform while it hadn't lost much of its value, and switch to a motherboard which will be supported until 2020.
Maybe he will buy the top 7nm zen3 later.

>>61087855
It's Haswell, 2012 architecture. Kill yourself, poojeet.

File: Girls.png (490KB, 449x401px) Image search: [Google]

490KB, 449x401px

>>61089464
>Literally has no idea that Specint2006 is the industry standard measure for scaling
>Vomits out his totally ignorant opinions anyway

>>61089406
Some amd engineer already said they know how to improve ipc another ~15%

>>61089610
See >>61089471

File: 1487608128808.gif (2MB, 540x300px) Image search: [Google]

2MB, 540x300px

>>61089584
>being this autistic

>>61089589
I never said Threadripper isn't going to be great, but Intel does already have parts with 16+ cores. Expensive as nothing else, of course, but price wasn't what I was arguing.

>>61089584
There wasn't a single Haswell CPU released in 2012 you dumbfuck. It's all 2013-2014.

>>61087111
>he got like 25% better multi-threading
why would he need 25% better multi-threading? what he does for a living?

>>61089709
Shills on 4chan from 8 different virtual machines.

File: 1493930001923.png (222KB, 1080x1920px) Image search: [Google]

222KB, 1080x1920px

>>61089709
>no one NEEDS multithread performance goy, buy our $300 quad cores

>>61087111
It's probably even less than that, considering his 5820K was overclocked to 4.7

>>61089760
>everyone NEEDS 10%-15% better multi-threading performance for 500 jewbucks

>>61086187
You upgraded a CPU from 2014 to a CPU from 2017, and got basically no extra performance besides MAYBE 15-20% extra multi threading.

Also, pretending your single core performance is the same is just retardedly silly.

I can't think of a single situation your "upgrade" makes economic sense. You simply WONT see a return on your investment over the 5820k. The math doesn't work in your favor at all.

>>61089709
>why would he need 25% better multi-threading? what he does for a living?

if he didn't need the multithreading performance he wouldn't be on a 6 core intel hedt chip in the first place

ryzen is far more competitive than the hedt chips especially the lower end hedt chips unless the intel chips were massively discounted in price

how is this so hard to understand

>>61089464
Sure, that's always a factor to consider. However, as we saw with the Blender tests, Threadripper halved Ryzen's times.

>>61089863
1700 costs $300 you lying shill

File: 1489666511813.jpg (51KB, 600x515px) Image search: [Google]

51KB, 600x515px

>>61089863
>50% better multithreading? What do you need 25% better multithreading for? 10% more multithreading is not worth it, just buy an i7

>>61089980
>motherboards are free

File: 1495604146174.jpg (28KB, 466x404px) Image search: [Google]

28KB, 466x404px

>>61090009
>HEDT intel chipsets are free

>>61090022
I got mine with intel retail edge.

Paid under $300 for my 5820k and motherboard.


You don't have family who works in the industry to get you good deals anon?

>>61090032
>Need family to get extra jewtel parts for free
This jewtel tradition need to die

File: 1498203343224.png (65KB, 620x326px) Image search: [Google]

65KB, 620x326px

Yeah, family. You are totally not an intel shill who got paid in hardware.
>$5 coupon for a raid1 key deposited in your account

>>61090022
That's not what he said. He only said it would have cost $500 to switch to a 1700X.

>>61090022
OP already owned both the 5820k and the X99 board.

So he sold them and bought Ryzen, probably spending another $150-250 out of pocket to do so, even after selling the intel stuff.

Wow amazing value for a little more multi threading performance that he'll likely never use.

>>61087068
AHAHA OH SHIT IT'S KEVIN MCGILL

>>61086187
coping

File: 1483587203666.png (17KB, 328x215px) Image search: [Google]

17KB, 328x215px

>>61090032
>Paid under $300 for my 5820k and motherboard.

File: TB2f4vIvrJmpuFjSZFwXXaE4VXa_2341477237_1.jpg (470KB, 1060x1569px) Image search: [Google]

470KB, 1060x1569px

>>61089670
Now pay attention, dumbfuck: That Xeon Gold is what the 18-core Skymeme-X is going to.be a rebrand of. The 16-core Threadripper is going to score ~1.99*1800x score, as in ~3,100-3,200 in Cinebench R15 for $999. Now fuck off with your dumpstered Skymeme-X space heaters.

File: 2017-04-22 11_20_33-CPU-Z.png (14KB, 401x398px) Image search: [Google]

14KB, 401x398px

>>61090093
yes?

>>61090076
such a good deal goy

Considering a build for minor gaming and doing video editing and 3d modeling. Wouldnt mind fucking around with linux ans gpu passthrough.

Ryzen seems a great option, but i can swing threadripper if reviews are good. One of the big attractions of am4 is building a system now and just upgrading gpu/cpu in 3-4 years. Does the am4 socket longevity also extend to threadripper?

>>61090501
>Does the am4 socket longevity also extend to threadripper?
Would assume so, they don't have any reason to switch sockets there.

>>61090501
Are you really sure you can utilize a Threadripper? 8C/16T is already quite a lot and few program are so truly parallel as to take full advantage of just that. It would be a pity to plow money into a Threadripper if Ryzen can serve your need equally well.

>>61090574
Unless he does a lot of pure CPU based rendering or similar, i simply can't see any reason to waste money on thread ripper as a normal consumer besides bragging rights.

>>61090543
AM4 - 2020. TR4- 2022?

>>61090574
>>61090588
Ill admit bragging rights has something to do with it. When I was younger and poorer I always wanted an sr2. I also have about 10 years of family video im going to have converted to digital (figuring out best way to do this) so for quite a while ill be working on that and would like to be able to 3dmodel/play games while stuff renders out.

I figure if the price difference is less than 5-700 bucks between a decent tr setup and a 1700 ill splurge a bit for stupid reasons.

Its only realy worth it to me if TR ends up having thebsame longevity. I upgrade every 4-5 years and pass on the old system to siblings or family friends so its financially I can justify that since id not have to buy new mobo and the like.

>buys i7 5820k
>turns right around and buys Ryzen 1700x
>thinks multi threading is important for gaming
Should have bought an 18 core for higher fps

File: 37526e9b79e14996251d509706ac58593.jpg (77KB, 1024x768px) Image search: [Google]

77KB, 1024x768px

>>61090649
forgot to post image so you know where i stand

>>61090645
As long as you understand that you likely wont be fully utilizing threadripper, it's not exactly a BAD buy.

In ~5-6 years when you upgrade platform again threadripper will make for a very beefy home server. 2-4 cores for PFsense, a core for DNS, a core for DHCP, a core or two for media server duties, etc, etc.

>>61090686
>a core for DNS, a core for DHCP
What kind of a home network are you running where DNS and DHCP use more than 1% of even an old Athlon XP core?

>>61090750
When you have 16 cores and 32 threads, why the fuck would you care about something that only takes up a single core?

>>61090645
>I figure if the price difference is less than 5-700 bucks between a decent tr setup and a 1700 ill splurge a bit for stupid reasons.
Sure, as long as you're aware. Threadripper might even have benefits for fewer-threaded programs too, given the extra memory channels and all, so it's not like it's a total waste or anything.

>Its only realy worth it to me if TR ends up having thebsame longevity.
I don't think there has been any official confirmation, but I'd be surprised if it doesn't. AMD has always been good about not switching platforms.

>>61090760
I'm not saying you should care that it does, I'm just wondering how you'd manage to make it do that.

>>61090686
>a core for DNS, a core for DHCP
Lmao. You've never run a home server.

>>61090797
Lmao, you've never run a home server with an over abundance of cores available.

ESXi with all your services in their own VMs, with each having their own cores (or more cores for higher load VMs)

>>61087068
>never understood attention seeking or validation seeking type behavior.

You use a trip.

>>61086457
When will it permanently occur to you that you are dead and bankrupt already?

Delete your trip and never return here.

>>61090953
>all your services in their own VMs
Utterly disgusting. Only MSCEs would configure things that way.
>with each having their own cores
Only control freaks think that's a good idea. Just let the host OS allocate CPU resources optimally. All you're doing by locking affinity is just withholding possibly free resources from another process/VM that could make use of them.

>>61091749
Having isolated VMs makes diagnosing issues dead fucking simple though.

If something is wrong with a particular VM, you know exactly what the issue is, either hardware associated with that VM (NICs, RAM, Hard drives, etc), or configuration issue with that VM.

Good luck trying to diagnose specific issues that might crop up 2 years later when you've got everything in the same OS running along with each other where you've got multiple drives, NICs, etc all being used by different things.

Sorry but no thanks

>>61091749
What is a vcore? Fucking shit you dont even know what you are talking about

>>61091830
Troubleshooting is for plebs and poorfags. Real sysadmins dump the entire hardware in the dumpster and order brand new racks from Intel(tm) at the first sign of trouble.

>>61091830
>hen you've got everything in the same OS running along with each other where you've got multiple drives, NICs, etc all being used by different things
This is exactly my situation, and I've never had a problem. If there's a hardware problem, dmesg points out exactly what piece of hardware is at fault. On the other hand, having everything is one OS image makes it far easier to share data that should be shared. Process virtualization is exactly what operating systems have done since the 60s, there's no point in having one process per VM.

>>61091858
>What is a vcore?
A VM core that the host OS can allocate to the optimal physical core unless you're stupid and explicitly bind it so as to prevent the host OS from doing its best.

>>61086187
The only meme is your belief that you were somehow CPU bottlenecked in fucking video games with a 4.7 GHz 5820k. Basically lmao@u r life

>>61091863
I still wouldn't bother with that shit since i'd be dedicating a good chunk of hardware specifically to Pfsense anyway. ESXi just makes the most sense.

If you wanna do windows server or some shit and let the OS do all your shit, fine. EXSi with Pfsense on some of the hardware and your windows server on the rest. Then configure your windows server to do whatever else you want.


I still would use ESXi though as my hyporvisor and run multiple VMs.

>>61086745
I've got a 5820k myself. If I hadn't bought last fall I'd heavily consider a 1700X.

>>61091909
>windows server
See, there's your problem. For some reason, it's always Windows admins that want to do stupid shit like that. I don't know if it's their shit OS that forces them to do stupid shit, or if it is that stupidity is infectious.

>>61086187
I would suck her dick

>>61091921
Sure, but no one who already owns the 5820k would be retarded enough to make that "upgrade" since for single core performance it most certainly is a downgrade, and for highly threaded situations it's still only ~20% faster.

>>61091940
I already told you I wouldn't be using windows in my own setup moron. I was using it as an example there of why someone might want to do such a thing.

>>61091967
Then you just make no sense, though. What do you even imagine you gain from doing something like that on Linux/BSD?

>>61092004
I already fucking said, ease of trouble shooting and keeping things segregated so if there IS an issue it's basically immediately identifiable.

It also insures you don't ever have any conflicting packages since everything is running in it's own environment.


Fuck, the WHOLE god damn point of thread ripper is being able to run several VMs all in their own encrypted work spaces each either their own encrypted RAM sets, and CPU cores.

>>61092029
>I already fucking said, ease of trouble shooting and keeping things segregated so if there IS an issue it's basically immediately identifiable.
Please explain in more detail how you imagine troubleshooting to be made easier by that. It's literally the same procedure either way, except you don't need to traverse multiple layers of VMs.

>conflicting packages
What? I hardly even consider what I'm installing on my machines, and I've never had such a trouble. Shared libraries are even explicitly versioned so as to not have conflicts.

>Fuck, the WHOLE god damn point of thread ripper is being able to run several VMs all in their own encrypted work spaces each either their own encrypted RAM sets, and CPU cores.
If you're a cloud provider, that's a great feature! More of a thing for EPYC, however. Otherwise, it's for simple, ordinary parallelism.

>>61092071
So please, tell me how YOU would use a single thread ripper CPU with 16 cores and 32 threads to provide PFsense routing capabilities, DHCP, DNS, media server (plex streaming or similar), and a work environment for CPU based video encoding.


Only one OS. Go

>>61092099
Just run the stuff, quite simply.

>>61092128
lol kay, guess we're done here.

>>61089709
reminder ryzen 5 and 7 now outperform the 7700k in rise of the tomb raider thanks to optimizations

Once you've properly multithreaded a videogame, AMD's processors catapult ahead of Intel. The only reason Intel has a "speed advantage" is entirely due to being the market leader everyone had to work around for the past 8 years. Everything is already optimized for Intel's lineup, virtually nothing is optimized for Ryzen. Once you change that, the 'speed advantage' Intel has drops to nothing or goes steadily in AMD's favor.

And that's not even taking into account superior frametimes and minimum framerates that few "reviewers" care to mention. The performance gap between AMD and Intel is going to increase in AMD's favor until 2020 - 2022 when Intel gets an actual new architecture to market instead of more rehashes. Only then will we, maybe, see some real competition from Intel.

Additional reminder that Zen 2 is coming on 7nm in 2H or 3rd quarter 2018 and will have 40% lower power consumption, be able to hit 5 Ghz, and continue to use the same AM4 socket everyone already has for Zen. Intel is completely FUCKED.

>>61092132
Why? What do you even imagine is wrong with that? Do you somehow think cores are getting withheld from you just because you're not running multiple OS images?

>>61092143
wouldn't any multi threading optimizations HELP the HEDT intel chips too though??

If tomb raider is running better on Ryzen because of multithreading improvments, 6-8 core intels should see a similar uptick right?

>>61092157

Who the fuck would recommend running EVERYTHING in a single OS? Especially something like a router(pfsense), where you probably want to be BSD based.

Installing pfsense and just installing some packages already optimized for pfsense is just dead simple compared to setting up your own router in BSD, setting up your own configurations, installing packages to monitor traffic, packet shaping, etc.

Why the fuck go through all that effort?

And if you think troubleshooting all this is the fucking same as what i'm suggesting, you're insane, or have never really done much router OS type shit.

>>61092201
I've never run PFSense, but if you have to set up a separate VM in order to be able to maintain or debug it, then the only possible conclusion is that iptables is superior. One rule for NAT, one sysctl to turn on routing, and you're done.

>>61092216
I have a 1gbps up and down connection, i need a bit more to my router OS than that, not to mention wanting decent VPN throughput with AES-NI encryption.

It's easier just to dedicate the hardware you need to the task and let it run in it's own user space segregated from everything else.

>>61092249
>i need a bit more to my router OS than that
But why? Even if you need more CPU time to route that, my 100/100 connection hasn't even blipped on CPU time. Even if it's traffic shaping you want, Ive' got that with ~50 lines of shell script to call from ifup, for a relatively complex, hierarchical setup.

>not to mention wanting decent VPN throughput with AES-NI encryption
Admittedly I haven't tried that, but I can't imagine it being more than 2-3 commands of IPSec setup. Is it just that you're incapable of reading manpages, and therefore need VMs with prebaked recipes?

>>61092299
You can't seriously expect that to be as optimized as pfsense which is used across the industry...right?

I severely doubt you had your 5820k at 4.7

>>61092312
Yes I can. Linux routing with qc QOS and iptables is also used across the industry. Also, as I said, with 100/100 uplink it hasn't even blipped in terms of CPU time, judging from which I could probably serve 10/10 Gbps with a single core.

>>61092332
>tc QOS*

>>61092332
I mean, if you seriously need 2-4 Zen cores to route 1 Gbps traffic, that seems to say a whole lot about pfsense optimization.

>>61092332
>judging from which I could probably serve 10/10 Gbps with a single core.

lololololol

with PFsense and a dell R210ii (quad core xeon) I can't even get much over 8gbps

>>61092365
It was an exaggeration, I'm just saying that it would take quite a lot to get into problems with 1 Gbps.

>>61092364
You could do pure routing with 1 core, but good luck doing any sort of live traffic analysis.

>>61092397
I've already got live traffic visualization using the data TC collects.

>>61092397
>>61092414
Either way, it's not like your routing gets quicker from running in a VM. It sounds like the only reason you need a VM is because you're too scrubby to set up things and just need a prebaked OS image to do basic stuff.

>>61092395
Sure, but you're seriously underestimating VPN throughput capabilities. To get 1gbps VPN throughput from a big name company (cisco, etc) you'd be looking at over $3,000 in hardware.

>>61092436
See >>61092431 though. It's not like your VPN gets quicker just from running in a VM.

>>61092436
>>61092457
Challenge accepted, though. I've been thinking about setting up VPN between a few private locations and a few hosted servers I manage lately. Just seeing how hard it is to set up and get throughput might be the last incentive I need to do that.

>>61092475
Let me know how that goes, right now my hardware is keeping me capped to about 350mbps throughput with 256AES.

>>61086187
Gaming? How old are you?

File: IMG_0011.png (23KB, 640x380px) Image search: [Google]

23KB, 640x380px

>>61092161
>If tomb raider is running better on Ryzen because of multithreading improvments,
Ryzen is still slower

>>61092489
Have you tried to pin down why, though? It can't be just cryptographic throughput, since with AES-NI, most semi-modern CPUs are capable of multiple gigabytes per second.

>>61092522
>are capable of multiple gigabytes per second

Depends what sort of hardware you're running.

Right now i'm using a C2558 (essentially a pfsense SG 4860) quad core atom CPU with AES-Ni support.

>>61086457
How are you still allowed to post here?

>>61092249
Aes-ni is an intel extension. Does amd support aes-ni?

I have three servers with e5-2670s, i'd like to tip my toes into the pool with threadripper if it supports proper vfio pci setups. Oh and 10gige, i want that too

>>61086832
FUCK YOU!!1 INTEL IS ######111111!1!111!!!!!

>>61094253
While the instruction set was proposed by intel, all AMD CPUs from the past 5-6 years support it as far as i'm aware, not sure why that'd change with threadripper So I assume it also has it.

>>61086745
Did this trip get hijacked or have you actually STILL BEEN ON HERE since 2008?

>>61094352
there was the thread the other day with the dude using his university cluster on trips. It's possible he decided to break some older ones.

>>61094556
that was a gud thred
he freed up a tonne of locked ones

>>61088749
>Pentium is an office lady processor

Not saying the Pentium is a real powerhouse, but I feel like the Celeron would be the best (cheapest) option for this application.