• Global community
    • Language:
      • Deutsch
      • English
      • Español
      • Français
      • Português
  • 日本語コミュニティ
    Dedicated community for Japanese speakers
  • 한국 커뮤니티
    Dedicated community for Korean speakers
Exit
0

Best PC workstation for CS5.5 Graphic Design Work

New Here ,
Oct 08, 2012 Oct 08, 2012

Copy link to clipboard

Copied

Hey all,

I'm a graphic designer looking to get a new system. Budget is about $3-$5k. I've been a Mac fan for years but want a PC this time; I can get near cost pricing on PC parts as husband is in IT.

Conscidering this system:

Lenovo D30 XEON E5-2609 Workstation

2nd processor: INTEL XEON 2.4G 2011 10M

NVIDIA QUADRO 4000 (2GB)

16GB ECC DDRIII RAM

Intel 520 Series SSD + WD Black 72k HD for storage

Windows 7 x64 Pro

Dual 24" 1080x1920 HD LED Samsung Monitors

Work I do:

- Graphic design - logos, websites, trade show displays, advertising, etc. (Mostly Illustrator)

- Light programming (Mostly Dreamweaver)

- Photo editing (Photoshop)

- Desktop Publishing (creation of magazine, advertising maps/guide booklets) in InDesign.

Often use large files; trade show displays are huge generally. Also like to use iTunes, Chrome, etc in the background.

Using latest CS6 suite

Hardware Questions:

- Am I better to put the money into one high end processor, or dual processors?

- Am I better to use workstation Xeon processors or Desktop i7 processors?

- ECC RAM or regular RAM? Is ECC worth the extra money?

- Video Card - is the Quadro 4000 worth the money, or am I better to go with something cheaper and spend more on RAM and processor, or go with two Quatros?

Thanks in advance for taking the time to give me a hand and help recommend what hardware to use

Views

19.1K

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 08, 2012 Oct 08, 2012

Copy link to clipboard

Copied

For apps you listed, unless you work with 500mp images in PS, a single CPU system seems a better fit.

Workstation-class dual CPU systems usually only make sense in two cases:

  1. when you can't achieve the performance you need with a single CPU.  E.g. CAD, 3D modeling apps, fluid dynamics, etc.
  2. when a feature you need / want that is only available on dual CPU motherboards and systems.  Such as memory expansion (to 512GB on C602 chipsets), I/O bandwidth (total number of PCIe lanes available / possible).

Another reason people may go for a workstation class system is the engineering, build quality and certain support aspects that come along with it.  It's not easy to find a desktop that has all of the features of, say, HP Z800 or Z820:

  • fully tool-less assembly including PSUs and optical drives
  • optimized airflow and reduced power consumption, heat generation and vibration that go along with it (not easy to achieve on a system put together from COTS components due to variable heat sink sizes).
  • handles
  • rackmounting
  • on-site support (3 years usually included, up to 5 years available, including an option of a 4-hour on-site response)
  • single point of support for all components sourced from that vendor.  E.g. all patches and updates are available on a support page dedicated to a specific workstation model.

Have I missed anything?

Yet another reason is vendor certification - e.g. Avid used to only certify HP Z-series systems and their support wouldn't talk to you (if at all) unless you have one of those.  It often helps to have a system that support is familiar with.

That said, if a single CPU (Core i7) system does what you need performance-wise, then a Xeon-based system can't touch it (BFTB-wise) with a mile-long stick.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Enthusiast ,
Oct 09, 2012 Oct 09, 2012

Copy link to clipboard

Copied

HI,

to answer your questions directly

Dual Xeons? almost never unless red 4k with multiple layers or heavy animation.

Quadro? almsot never unless using solid works or have a need for 10bit color (and a cheap blackmagic card takes care of this) and are based on older GTX cards in other words you pay far more for slower tech

ECC ram NEVER NEVER on a workstation only for a server with high i/o transactions its much slower.

a single i7 with faster GHz than Xeon will vastly outperform and be much less money..

usually buying HP or other tier 1 means giving up options or having to spend way too much to get them... why not have your hubby build it?

Scott

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 09, 2012 Oct 09, 2012

Copy link to clipboard

Copied

Scott Chichelli wrote:

Quadro? almsot never unless using solid works or have a need for 10bit color (and a cheap blackmagic card takes care of this)

Scott,

1. Which "cheap blackmagic card" takes care of 10-bit color on a display monitor?  If none, as I suspect, how do you expect that "cheap blackmagic card" to take care of 10-bit color in Photoshop, for instance?

2. Do you know if any GeForce card that is covered, or has an option to be covered, by an on-site replacement warranty (that comes standard on all components on HP workstations)?

Thanks.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
New Here ,
Oct 09, 2012 Oct 09, 2012

Copy link to clipboard

Copied

Yeah building a system is the other option. He can do that, the reason we were looking into the Lenovos is he gets 50% off workstations, and another 35% off the options (like the high end graphics cards) as with building a system he will get equivelant to New Egg pricing; maybe 2-3% off if he is lucky.

As for warranty not a huge concern; husband can handle all that.

Any suggestions for base Lenovo systems? Or perhaps intel and a custom build it the way to go. For $290 he can do 64GB of RAM in the custom built one.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 09, 2012 Oct 09, 2012

Copy link to clipboard

Copied

Any suggestions for base Lenovo systems? Or perhaps intel and a custom build it the way to go. For $290 he can do 64GB of RAM in the custom built one.

CPU: choose a 6-core (1333MHz base clock) or an 8-core (1600MHz base clock) CPU rather than the 4-core 2609 running at 1066MHz base clock: faster memory (along with faster CPU) translates to a significant performance boost, and not all too expensively: a 6-core E5-2620 is $250 on top of a 4-core 2609.

Memory: I'd install registered memory rather than unbuffered.  E.g. install 4x8GB ECC reg up front (per CPU) - so that if you need to upgrade the memory in the future, you could.  Unbuffered memory limits upgrade choices; is less reliable than registered.  Memory upgrades are the most common, popular system upgrades, so if there is one thing not to limit yourself on, it's memory.  It's entirely possible PhotoShop CS8 will run smoother on 128GB vs. 64GB.

Display monitor.  PhotoShop loves screen real estate, and dual 1080p monitors... are OK but wouldn't give you the joy, the freedom of working on a large glass like a 2560x1440 or 2560x1600 monitor.  I work with hi-res stills from 5D and d300, and my 30-incher is simply irreplaceable; going back to a 24" monitor (however many) is just not the same thing.

Most of these monitors are a grand or more; HP however has a 2560x1440 IPS 27-incher for under $800.

The rest of your system looks quite all right.   (If this will be your main work tool for the next few years, pinching a few dollars here and there will make you feel good in the short term... Will it in the long one?)

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 09, 2012 Oct 09, 2012

Copy link to clipboard

Copied

Memory: I'd install registered memory rather than unbuffered.

I'll have to take that back: unbuffered memory is OK, at least according to Lenovo. Unlike HP that doesn't recommend fully populating both banks with unbuffered memory, Lenovo seems OK with that.

Just keep in mind the maximum size of an unbuffered DIMM is 8GB; registered - 32GB. Maximum unbuffered memory with a single CPU on D20 (or HP Z820, or Dell T7600) - 32GB, dual CPUs - twice that.  If you do foresee a possibility of using more than 64GB RAM in the next few years, use registered RAM.

HTH.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
New Here ,
Oct 11, 2012 Oct 11, 2012

Copy link to clipboard

Copied

Thanks for all the helpful advice!

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guru ,
Oct 12, 2012 Oct 12, 2012

Copy link to clipboard

Copied

Keep in mind single Xeon systems give no advantage over X79 workstations since the CPU's are locked. You can get 8 Core Xeon chips versus the 6 Cores X79 CPU's but the X79 CPU's over greater clockspeed options which is far better for the Hardware MPE engine. If you are looking at the Xeon route then the only configuration that makes it worth the investment is Dual Xeon setups. Obviously some configurations offer you the ability to Add a 2nd Xeon later if the budget is limited now.If that is not in your plan then I suggest you look at the X79 Workstation platforms instead.


Eric

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 12, 2012 Oct 12, 2012

Copy link to clipboard

Copied

Keep in mind single Xeon systems give no advantage over X79 workstations since the CPU's are locked.

What about bus speed (8 vs. 5 GT/s), memory expansion, ECC, PCIe 3.0, four times the number of PCIe lanes supported (important for people with Kona 3, ARC-1882, and a 10GbE card)?

I am not talking about what they all mean in real world, just the "no advantage" part.  Perhaps we should re-think it a bit?

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Enthusiast ,
Oct 12, 2012 Oct 12, 2012

Copy link to clipboard

Copied

ECC=BAD and SLOW.. not a benefit but a flaw.. when we build dual Xeons we don’t use that garbage either ECC is for high transaction I/O servers NOT workstation..

I can do 64gig on a an X79 anyone needing more than that will be buying a DUAL Xeon…

X79 has PCIe3.0

More PCIe lanes? On X79 I have more than enough for a BM card or Aja(almost never AJA ) and a Raid card and a Pro audio interface

Bus speed? ::::sigh:::: Eric your turn..

Scott

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 12, 2012 Oct 12, 2012

Copy link to clipboard

Copied

You personally not ever needing more than 64GB RAM ever - somehow makes it true for the rest of the editing, compositing, 3D modeling content creation world?

Intel® Core™ i7-3930K Processor:

- PCI Express Revision: 2.0

- # of PCI Express Lanes: 10

What are the numbers for E5-2600 series?

(I think I asked you before and never got a response. Ever tried fact checking and doing a little research before posting publicly?)

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guru ,
Oct 12, 2012 Oct 12, 2012

Copy link to clipboard

Copied

You might want to update yourself on the specs of the X79 chipset. Please read the Intel Docs below:

http://www.intel.com/content/www/us/en/chipsets/performance-chipsets/x79-express-chipset.html

The X79 platform has as many PCi-E lanes as a single E5 Xeon. Remember Both Xeon CPU's have to be populated to get all of the PCi-E lanes. Half of the lanes are not active if you only populate 1 CPU. The ram is also halved with only 1 CPU populated.

I believe you can ask many DIY editors here how their PCI-E 3 raid cards are running in their X79 systems along with I/O cards.

The QPI speed is a point however, the X79 White paper lists PCi-E Gen 2 instead of 3 when it refers to QPi speed of 5GT/s. Since the X79 boards are PCi-E Gen 3 now I am not sure if that is still accurate. However I will have to ask the Intel engineers since you brought it up. Regardless though the performance on the Single Xeon and Even the Dual Xeon has not shown any improvement over the X79 in Premiere even with the possible difference in QPI speed.

Eric

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 15, 2012 Oct 15, 2012

Copy link to clipboard

Copied

Eric,

Aside from trying to educate me ... Where in those X79 docs on Intel web site does it say PCIe 3.0?  Must have missed it somehow.  Looking at specific X79 boards - sure, they support PCIe 3.0, but the X79 chipset per se - doesn't seem to, at least not according to Intel.

The X79 platform has as many PCi-E lanes as a single E5 Xeon.

That's what I am trying to reconcile, that i7-3930K specs show 10 lanes of PCIe 2.0 yet the X79 chipset shows 40 lanes... also of PCIe 2.0. Yet motherboards appears to have 40 lanes of PCIe 3.0.  Do you know what's up with that?

Yet on another point: even if today 64GB RAM is a lot, in 3 years it may not be - RAM usually is the number one upgrade on older systems.  Wouldn't that 64GB limitation hinder people needing upgrade in a few years?  In other words, isn't that at least a minor advantage of Xeons?

Why is ECC bad and slow?  Would be great to see independent and trusted opinions on that - this is the first time I hear it.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guru ,
Oct 16, 2012 Oct 16, 2012

Copy link to clipboard

Copied

Thee Intel Diagram and papers are likely the original spec and have not been updated with the current manufacturing. This is not uncommon. Remember the Specification by Intel are minimum spec requirements. The System Board partners can change them as long as the minimum requirements are met. In this case All the X79 boards I have seen support PCI-E Gen 3 and also have 40 lanes to work with. That is within the spec Intel lists.

64GB of ram is more than enough for editing with the current codecs. Very few production require more than that right now. By the time that changes in the future, likely the hardware will have changed to a new platform.

ECC is slower than unbuffered because of the added latency for ram instructions due to the parity check. I have not seen a study on this effect since the DDR2 and DDR1 days but the overall adding of latency over a period of time is significant for a performance workstation. Server operations though are so critical especially with Database integrity that ECC is absolutely required for those.

Eric

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 16, 2012 Oct 16, 2012

Copy link to clipboard

Copied

Thanks Eric,

Remember the Specification by Intel are minimum spec requirements.

So the 64GB RAM limit can be exceeded at will?

Sorry, the diagram doesn't look like the "minimum spec" that can be exceeded at will, and I'd love to see what Intel (or anyone else) has to say about performance differences between native on-chip PCIe 3.0 QPI support vs. that on the chipset that seems to be added on, and is not natively supported by the CPU.  That was my original question.

ECC is slower than unbuffered because of the added latency for ram instructions due to the parity check. I have not seen a study on this effect since the DDR2 and DDR1 days but the overall adding of latency over a period of time is significant for a performance workstation.

ECC slower than unbuffered? You meant, non-ECC, perhaps?  Unbuffered has very little to do with ECC.  If you meant unbuffered vs. registered: according to this, registered can be some 5% faster than unbuffered on 5520 platforms with more than two DIMMs per channel.

The point here is, Scott doesn't say "slower" he says "slow" (and "expensive").  From what I am seeing, speed difference is 2% at worst.  That does not qualify as "slow", only as "insignificantly slower".  It's not "expensive" either, it "somewhat more expensive" - and the difference in price is usually vastly minuscule next to the total system price.

So perhaps "slow and expensive" with respect to ECC needs to be re-phrased into "marginally slower and a little more expensive", or not mentioned at all?

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guru ,
Oct 17, 2012 Oct 17, 2012

Copy link to clipboard

Copied

The 64GB ram limit has to do with the memory controller in the CPU. That part/component is the deciding factor for the ram configuration. I have no idea why you reference this as part of the debate since none of Intel's partners have any input into this.

Please reference the Block Diagram below:

http://www.intel.com/content/www/us/en/chipsets/performance-chipsets/x79-express-chipset.html

Do you notice the part where is says Multicard configurations? That means the motherboard partners can change that as will. However the X79 Chipset that Intel manufacturers (ie not 3rd party which means the PCi-E controllers are native) supports up to 40 Lanes of PCI-E bandwidth. The 10 lanes listed under the CPU refers to the PCI-E lanes used for the DMI bus and other direct connects to the CPU if required.

Please reference the link below for Intel's X79 board designed by their engineers.:

http://www.intel.com/content/www/us/en/motherboards/desktop-motherboards/desktop-board-dx79si.html

Notice that even their board specifies PCI-E Gen 3? If Intel builds their boards with Gen 3, can you not see that once again the specs listed for X79 chipset is Minimum/recommended specs. Also at the bottom of each page on the X79 Intel states they can change the specs without notification and at will. That means they can do just that. Why do you think the amount of Sata ports on the Intel controller on the X79 board changes based on model. Intel specifies how many Sata connections their Controller will handle and then leave it up to the 3rd party partners to decide what they will use.

I meant ECC and your comments are incorrect. ECC has everything to do with this because Registered Sticks are almost always ECC. ECC ram adds the latency to the instructions period. This has been documented for years. You might want to also read the entire description of Registered which aslo explains why normally Registered ram is slower than Unbuffered because it's always a cycle behind. The reason is there is another controller with the registered ram that has to process the instructions. This once again add's latency. Add to this that Registered Dimms are normally slower GHz than Unbuffered, the overall bandwidth of unbuffered is higher. As a side note. You can change the 2N for the command rate on unbuffered if the system is not overclocked and the ram is higher grade. Their comments here regarding the command rate are simply out dated. I have tested the overall bandwidth of the ram with 1N command rate and 2 N and the difference was so marginal, it was not even worth the discussion. However the difference in performance overclocking the CPU was so significant that having a Xeon system with Registered ram was significantly slower even with 8 Cores that even an I7 Gen 990X system beat it in benchmarking with Adobe. So you may want to test the configurations before you assume.

Eric

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 17, 2012 Oct 17, 2012

Copy link to clipboard

Copied

Eric,

TL;DR, with the exception of "your comments are incorrect" and "All your ship belong to us. Resistance is futile. You will be assimilated." OK maybe not the 2nd quote.

Do you have anything to back up your declaration of ECC being "slow and bad" (how slow? how bad? got sources?) - and maybe only then write something like "your comments are incorrect"?

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Guru ,
Oct 18, 2012 Oct 18, 2012

Copy link to clipboard

Copied

Example 1: From HP oddly enough and this is abour RDIMM itself without accounting for ECC on page 2.

http://h20000.www2.hp.com/bc/docs/support/SupportManual/c02126499/c02126499.pdf

http://en.wikipedia.org/wiki/ECC_memory

Reasons for not using ECC:

*A performance decrease of around 2–3 percent, depending on application, due to the additional time needed for ECC memory controllers to perform error checking.

http://www.pcguide.com/ref/ram/errECC-c.html

*Unlike parity checking, ECC will cause a slight slowdown in system operation. The reason is that the ECC algorithm is more complicated, and a bit of time must be allowed for ECC to correct any detected errors. The penalty is usually one extra wait state per memory read. This translates in most cases to a real world decrease in performance of approximately 2-3%

http://www.crucial.com/kb/answer.aspx?qid=3692

*Using ECC decreases your computer's performance by about 2 percent. Current technology DRAM is very stable, and memory errors are rare, so unless you have a need for ECC, you are better served with non-parity (non-ECC) memory (note - Keep in mind this is a memory manufacturer that makes more from ECC Modules than Unbuffered modules)

I will see if I can find some more official links for you when I have more time

Eric

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Enthusiast ,
Oct 18, 2012 Oct 18, 2012

Copy link to clipboard

Copied

i think the ECC % reduction in performance is greater for Audio/video. particularly in projects that are more ram intensive like AE or other animation..

i say we get ahold of Top Power and have them send us 64 gig ECC (loaner) so we can do a benchmark and put this to bed once and for all..

you know me i dont rely on ANY 3rd party source i have to see it with my own eyes...

Scott

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 18, 2012 Oct 18, 2012

Copy link to clipboard

Copied

i think the ECC % reduction in performance is greater for Audio/video.

Scott, I think it's a great idea not to get into speculations until there is solid proof.  Just like you said.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Enthusiast ,
Oct 19, 2012 Oct 19, 2012

Copy link to clipboard

Copied

assuming its only 2% which i know its more.. (way more for animation) but lets say 2%.

and to quote you from the "volumes" thread loosely  "other than being able to beat a benchmark"  or something like that..

lets say i get 3% increase from the windows tweaks i do, then a little more from the bios tweak..

then another 2-3% on ram and then another 2-3% on disks..

well my friend we are in the 10+% mark now arent we? 

now lets say you edit all day long for a living. 5 1/2 days a week (typical of most small biz owners)

thats 286 days a yr.  10% is 28 days.. a month of time..

see thats the difference with me since day 1 i have always been about extreme power and increasing it any way possible.. (as long as its stable)

and yes my days began(in 97) as a tweaker/gamer/over clocker..

carrying that over into the professional market has paid off..

lets put that another way.. how would you like to have back all the CC fees for a yr.. after all thats only 3% ish..

Scott

ADK

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 19, 2012 Oct 19, 2012

Copy link to clipboard

Copied

Your accumulated credit card interest makes ECC "slow and expensive"?

Given that a lot of editing is not encoding or rendering (media management, ingest, playback, storyboarding), what impact ECC's 2% performance delta will have at the end of the day?

How about we get back to the original idea that until there's solid statistical proof across multiple platforms and memory types, we stay away from speculation, irrelevant parallels, an statements like "ECC is slow and expensive"?  Because so far the evidence is pointing the other way.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Enthusiast ,
Oct 19, 2012 Oct 19, 2012

Copy link to clipboard

Copied

WOW that’s it? Oh nevermind..

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines
Contributor ,
Oct 19, 2012 Oct 19, 2012

Copy link to clipboard

Copied

Scott, sorry - that came out wrong and I apologize.  It was confrontational, uncalled for, unnecessary.

I do hear you, I just wanted to stay on the message and on topic.  In this case, the impact of ECC's "slowness" on general system performance and editing workflow.

Tuning the system and maximizing its performance for a specific purpose is the name of the game and you guys are good at it.  If you say you prefer non-ECC to ECC on any given day, has weight to it already given your reputation.  Saying that ECC is "slow and expensive" has a whole different flavor to it though, especially coming from a reputable source.

My discussions with you and Eric do get confrontational - probably in part because I feel that you and Eric often make absolutist take-no-prisoners statements like "ECC is slow and expensive", "you don't need a video I/O card for monitoring any more", "Quadro is a waste of money", "HP is a waste of money", etc.

One thing is when statements like that come from a random user. Totally other when they come from people who carry a lot of weight on this forum.  I felt the need to question some of these statements - and well, here we go, the discussions often turn quite adversarial.

Perhaps there is a way to turn the heat down a bit, I'll work on it.

Votes

Translate

Translate

Report

Report
Community guidelines
Be kind and respectful, give credit to the original source of content, and search for duplicates before posting. Learn more
community guidelines