Sign in with
Sign up | Sign in
Your question
Solved

5850+phenom 9550=bottleneck?

Tags:
  • Graphics Cards
  • Bottleneck
  • Phenom
  • Graphics
  • Product
Last response: in Graphics & Displays
Share
October 6, 2009 8:56:00 PM

i currently have a system with a 4850 graphics card and a slow phenom x4 9550 (2.2ghz) i was going to originally upgrade to a i5 system and then another 4850 in CF but i wondered if i just skipped the cpu upgrade and went straight to a single 5850 (not now, like in Jan 2010).

my question is this:
would there be a huge bottleneck from my 2.2ghz phenom? and if so, how big? this idea might sound stupid but i just dont want to buy a new motherboard and cpu as that would be painful

More about : 5850 phenom 9550 bottleneck

Best solution

a c 176 U Graphics card
October 6, 2009 9:14:50 PM

I would say it would depend on if your overclocked, but seeing as the first phenom was such a dog at overclocking I don't think it would matter. I would say buy it and carry it over to the new build, but that doesn't make sense either seeing as the 5850 should be cheaper by the time your ready for it.

If I were you, I'd keep the system you have, and start buying the new system. If you are gaming only, you might want to consider a slow methodical upgrade. We don't know what your mobo is, but if you upgrade to a AM2+, you can use your current CPU and memory. Next upgrade would be the CPU, a nice 940 or 955, etc. After that, you should be able to support the 5850/5870. You'll be able to add each piece in and use it immediately, so its pretty painless to do it this way. The one down side is you'll still be stuck with AM2+ and DDR2. If you have the self control, saving up for i5 or AM3 if your on a budget would be better.
Share
a b U Graphics card
October 6, 2009 9:59:50 PM

Depending on your current mobo (mainly if its an AM2+) you might be able to upgrade directly to an AM3 quad core. You can check the manufacturer's website, or if you don't feel comfortable with that, post the model here. You can find this if you download cpuz and check the motherboard tab.
http://www.cpuid.com/cpuz.php
m
0
l
Related resources
a b U Graphics card
October 7, 2009 1:40:59 AM

What resolution are you playing at? The reason I'm asking is because at higher resolutions the gpu tends to do more work so at a high resolution the bottleneck could be a bit minimal.

Have you tried overclocking?
m
0
l
October 11, 2009 6:40:23 PM

I have a 9550 X4 BE, a BFG PSU 750W, a Saphire HD 4870, and 4G Ram, need to upgrade my HDD (fried my Caviar Black), as it stands, my system dosen't bottleneck, and I have a 740G Gygabit MoBo (AM2+)

I have played most games that I play at top res (x4 Bloom, Max anti Aliasing, etc) and been more than fine, BUT I am using 2x 1680x1050 Monitors, and am very happy. It games fine for what I play (Aion, WoW, L4F, TF2, Etc) and has rocked it out like a champ, BUT I am also doing the upgrade slowly method. My next priority is HDD, need to go Raid (3x Caviar Blacks for my budget). GL and good hunting.
Yes I am sure I do not bottleneck, but my system is prolly over kill for the games I play. Maybe not, IDK TBH.
m
0
l
a b U Graphics card
October 11, 2009 6:52:53 PM

You don't need to OC it a lot. Try a minor OC, and OC it 2.5/2.6 or such.
m
0
l
October 12, 2009 1:40:34 AM

ripperzane said:
...need to go Raid...

Why? Just curious, as RAID can be more of a hassle than what it's worth in performance.
m
0
l
a b U Graphics card
October 12, 2009 2:28:44 AM

I have a 160GB 8 m/s HDD that has a vista score of 5.7
I have a 320GB 16 m/s HDD (samsung) that has a Windows 7 score of 5.9
I also have 2x320GB 16 m/s HDD (seagate) in RAID 0 with a Windows 7 score of 5.9

I don't know why 2 HDDs in RAID 0 didn't improve my W7 score... >_<
m
0
l
a b U Graphics card
October 12, 2009 3:34:56 AM

Vista scores really don't mean anything when it comes to real-world performance. You could try some HD benchmarking utilities, such as HDTune.
m
0
l
a b U Graphics card
October 12, 2009 5:37:33 AM

I meant windows 7 score, not vista (which maxes out at 5.9)

Hmm, do you personally use HDTune?

What are some of your results for 7200 RPM HDDs with 16mb cache?
m
0
l
a b U Graphics card
October 12, 2009 6:29:48 AM

Yes, I do use it.

WD Black 640 GB (this is actually 32 MB cache, as I currently don't have any 16 MB cache HD's)

Max- 111.2 MB/sec
Min- 58.3 MB/sec
Avg.- 91.5 MB/sec

Access Time 12.2 ms

Burst Rate 126.7 MB/sec
m
0
l
a c 176 U Graphics card
October 12, 2009 10:28:33 PM

When he said use HDTune, he meant to do work, not benchmark your drives. If you test with AID0 and single drive mode for most tasks, there won't be much difference. AID0 will be a bit faster, but just like in this case spending the extra $70 on a faster CPU will improve everything more then adding the second drive.
m
0
l
a b U Graphics card
October 12, 2009 11:50:03 PM

I was assuming he meant benchmark, since he said he was using Windows 7 experience scores.

In addition to a faster CPU, more RAM would also help with whatever you're doing.
m
0
l
October 13, 2009 1:36:49 AM

@ the guy who said a 2.2 would bottleneck a 4670... It doesn't. Check out my post 'system balance' in the general homebuilt section. It has a 4600+ which is a dual core at 2.4 gigahertz. I mean it's possible that that 200 megahertz made a difference, but this is a quad core, which should matter for newer games.

The 4600+ did however bottleneck my 4850, so it should bottleneck the 4870 as well, and of course anything higher.
m
0
l
a b U Graphics card
October 13, 2009 2:33:12 AM

4745454b said:
When he said use HDTune, he meant to do work, not benchmark your drives. If you test with AID0 and single drive mode for most tasks, there won't be much difference. AID0 will be a bit faster, but just like in this case spending the extra $70 on a faster CPU will improve everything more then adding the second drive.

pepperman said:
I was assuming he meant benchmark, since he said he was using Windows 7 experience scores.
In addition to a faster CPU, more RAM would also help with whatever you're doing.

A lil of both I suppose.

I believe it did boost my HDD speed in "certain situations."
I did notice that I can transfer large files faster with RAID0...ranging from 5 to 70 mb/s, whereas with a single harddrive I usually only get between 3-40 mb/s depending on the location. But I don't notice any significant improvement in regular applications.

I do want to know the benchmarks since it'll give me a solid number to compare. At this point, I have an i7 920 and 6GB DDR3 RAM, so I can't really do much to improve that my CPU or RAM. I was thinking my original single HDD was the main thing holding me back which is why I did RAID. (I heard velociraptors were terribly loud and SSDs were pricey)
m
0
l
a b U Graphics card
October 13, 2009 3:24:06 AM

If you don't want to buy SSDs, then RAID is the best way to accomplish what your looking for. However, if your drives are more than a couple of years old, you might want to consider upgrading, as HDDs are always improving in terms of speed. You might want to look at the Samsung F3's, which beat Raptors in some benchmarks.

http://www.newegg.com/Product/Product.aspx?Item=N82E168...
m
0
l
a b U Graphics card
October 13, 2009 4:33:48 AM

^ Yep, I actually have the F1 320GB version.

I compared my F1 samsung to my 2xSeagates in RAID0...for some reason
it says my transfer min/max was higher for the single HDD? O_o

the F1 samsung got 70 mb/s min and 110 mb/s max whereas my RAID0 seagates got 70 mb/s min and 105 mb/s max... >_<
m
0
l
a b U Graphics card
October 13, 2009 5:13:18 AM

Depending on the model and age of the Seagates, that might make more sense than you think: the F1 series is a fast HD, and if I remember correctly, you said the Seagates only had 8 mb cache, while the samsung has 16 mb.
m
0
l
October 13, 2009 1:30:14 PM

trkorecky said:
Why? Just curious, as RAID can be more of a hassle than what it's worth in performance.

The discussions below my original post show there is some performance, but I am unsure of if that warrants the cost. My chief thought is redundancy, and I would likely be running a method of RAID that would be aimed at redundancy over performance, but that is a separate issue of research I need to conduct. I had a caviar black and a faulty PSU helped fry my old rig. I am aware that a bad PSU may potentially take the system regardless, but my budget for HDDs is able to handle either a raptor of smaller capacity or a couple (perhaps 3) caviar blacks for about the same ballpark (being $300 or less). The aim is to build a system that will not need any MAJOR upgrades for a little while (being hopefully over a year).

I currently have a mish mash of drives now, and it doesn't seem to be efficient, but it was what I had on hand at the time of building.
m
0
l
a c 176 U Graphics card
October 13, 2009 6:55:47 PM

Except AID0 doesn't have any redundancy, which is why I drop the R normally seen with the term. If your using a mix of drives, upgrading to one or two current ones can see nice speed gains seeing as you aren't using any older drives anymore. Trying to copy a file from an old 45MB avg transfer 120GB IDE drive to your new shiny 90MB drive is a good way to slow down the new drive. It will feel slow because its dealing with your old drive. I would say we need to know system specs and what you use the system for if you need to know what to upgrade. You aren't the OP, so we don't know if you have a Quad or not.
m
0
l
October 14, 2009 8:56:29 PM

Lol, I think it's funny how this thread kept going when I forgot all about it.

Anyway, thx for all the answers I was just curious about the limitations of a cpu. I used to think that the graphics card is the absolute king, but I guess not. I did an improvised test to see how close I am to bottleneck with my current 4850. I played CoD4 at 1440*900, 4AA, max AS, all settings maxed, extra texture detail and I got a wildely fluctuating FPS count: from ~70 up to 160. Then, I turned all settings off and played at 640*480, and the FPS was in the 300s so I figured I had quite a lot of headroom left. :ange: 
m
0
l
a b U Graphics card
October 15, 2009 2:08:59 AM

pepperman said:
Depending on the model and age of the Seagates, that might make more sense than you think: the F1 series is a fast HD, and if I remember correctly, you said the Seagates only had 8 mb cache, while the samsung has 16 mb.


They're both 16mb cached. Oh well... =/
m
0
l
a c 176 U Graphics card
October 15, 2009 6:37:11 AM

Not all 16mb cache drives are created equal. If the aerial density is higher on the F1, it would transfer faster.
m
0
l
October 15, 2009 2:36:47 PM

ripperzane said:
The discussions below my original post show there is some performance, but I am unsure of if that warrants the cost. My chief thought is redundancy, and I would likely be running a method of RAID that would be aimed at redundancy over performance, but that is a separate issue of research I need to conduct. I had a caviar black and a faulty PSU helped fry my old rig. I am aware that a bad PSU may potentially take the system regardless, but my budget for HDDs is able to handle either a raptor of smaller capacity or a couple (perhaps 3) caviar blacks for about the same ballpark (being $300 or less). The aim is to build a system that will not need any MAJOR upgrades for a little while (being hopefully over a year).

I currently have a mish mash of drives now, and it doesn't seem to be efficient, but it was what I had on hand at the time of building.

Redundancy is a great reason, I assumed you'd hit up AID 0 like most on this forum.

As for the guy posting about his hard drive score in Windows 7, I'm pretty sure they adjusted everything so it would only go higher than a certain value only if you're running SSDs. Pretty disappointing, yes, but that's how it goes. I'll ask my boss, she if she knows any better information or could lead me to someone else who does (I work for MS).
m
0
l
!