Alienware Graphics Amplifier CUDA-Z Speed
 
Notifications
Clear all

Alienware Graphics Amplifier CUDA-Z Speed  

 of  4
  RSS

jim_survak
(@jim_survak)
Eminent Member
Joined: 3 years ago
 
Posted by: nando4
@jim_survak, a Alienware 13 R1 uses a i5-75xxxU CPU. Their PCH PCIe ports max out at Gen2 speed, so the link would be x4 2.0. You can confirm this by reviewing the hwinfo64 section as shown by @itsage for his Alienware 13 R2.
Not that it makes a difference but I've got the very first Alienware 13 w/ AGA port, which specifically is an i5-4210U.
 
@itsage I powered off my Alienware when I went to bed last night so I can't remote into it now but I'll check it when I'm home for you to let you know what it says.
 
Posted by: itsage

If we take Dell's marketing at face value, the AGA performance should be closer to desktop 4x PCIe 3.0 and much faster than Thunderbolt 3 eGPU. This is not what I'm seeing.

Can't disagree with your findings.

 

I ran a test of Gears of War 4 with my AGA & my old desktop.

 

GTX 1060 6GB

Custom desktop, Core i7-3770K

GTX 1060 6GB

Alienware 13 w/ AGA

Avg GPU Framerate 60.2 49.3
Avg CPU Framerate (game) 167.1 80.4
Avg CPU Framerate (render) 148.8 44.9
Avg min Framerate 50.7 21.4

CPU frames are 2x-3x time that of the AW13's much slower CPU and, I assume, PCIe bus. My old desktop had PCIe x16 2.0

2018 Mac mini: Core i3, 8GB RAM, Sonnet 350 (aftermarket 550w PSU)+XFX Radeon RX-480 8GB Black Edition
2012 Mac mini: Core i7, 16GB RAM, Toshiba 1TB SSD, Seagate 1TB HDD
Custom: Ryzen 7 1700, 16GB Corasir DDR4-3200MHz RAM, 2x Corsair 500GB Neutron SSD, Seagate 3TB SSD, EVGA Nvidia 980Ti
YT:


itsage liked
ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 3 years ago
 

You are comparing apple to oranges in both cases scenarios. So it is hard to say if there is a performance bus issue.

Trying to compare the output performance of a bus when using a dual core, four thread, against a quad core, eighth treads can be misleading, as i already said multiple times!!

You can't simply compare the frame rate because obviously the the i5/i7 XXXX U can't compete with a true quad i5/i7 XXXX HK/QM/K when benched under heavy gaming applications.

Simply because basically, the capability of the cpu to spit out fps is halved. So you could apply this to each thing/task requiring cpu computation.

To conclude, it is true the Cuda-z numbers are a bit strange and intriguing. Cuda-Z need to himself trigger the full link speed. If using the tool along another program running will obviously return lower H2D bandwidth. But in any case you can't compare two different platform against with only few games, and pretend to give a proper conclusion.

The best would be to compare the 13 R2 against another laptop/macbook with an XXXXU cpu and a TB2/TB3 port.

2012 13-inch Dell Latitude E6320 + R9 [email protected] (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo Thinkpad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


ReplyQuote
Sky11
(@sky11)
Reputable Member
Joined: 3 years ago
 

Older Alienware R13 had Gen2 PCIe 4 lanes (Haswell and Broadwell versions if I am not mistaken)

 

To do: Create my signature with system and expected eGPU configuration information to give context to my posts. I have no builds.

.

ReplyQuote
itsage
(@itsage)
Illustrious Member Admin
Joined: 4 years ago
 

@wimpzilla, I agree with your assessment. What i'm trying to determine is whether the AGA is performing at the level Dell/Alienware claims. The host laptop, Alienware 13 R2 has both the AGA connector and Thunderbolt 3. However, the Thunderbolt 3 in this laptop is running at 2x PCIe 3.0.

After Dell's Thunderbolt 3 "half-truth" episode, seeing the CUDA-Z numbers from this AGA made me suspicious. Here are the numbers when I ran both the Razer Core and AGA through this Alienware 13 R2. AGA @ x4 should be much faster than Core @ x2.

Alienware 13 R2 Razer Core Internal AGA Internal Razer Core External AGA External
         
Unigine Valley 61.4 FPS 64.0 FPS 64.1 FPS 72.9 FPS
Unigine Heaven 60.0 FPS 73.3 FPS 71.3 FPS 76.7 FPS
Unigine Superposition 72.9 FPS 78.4 FPS 82.5 FPS 86.6 FPS
3DMark Time Spy 29.7 FPS 33.2 FPS 33.4 FPS 35.1 FPS
3DMark Fire Strike 60.1 FPS 67.4 FPS 70.9 FPS 84.1 FPS
         
Rise of the Tomb Raider 51.3 FPS 45.0 FPS 58.2 FPS 59.6 FPS
Tom Clancy's GhostRecon 40.4 FPS 50.8 FPS 52.1 FPS 53.8 FPS

external graphics card builds
best laptops for external GPU
eGPU enclosure buyer's guide

 
2020 13" MacBook Pro [10th,4C,G] + RTX 2080 Ti @ 32Gbps-TB3 (AORUS Gaming Box) + Win10 2004 [build link]  


ReplyQuote
jim_survak
(@jim_survak)
Eminent Member
Joined: 3 years ago
 
Posted by: wimpzilla

Trying to compare the output performance of a bus when using a dual core, four thread, against a quad core, eighth treads can be misleading, as i already said multiple times!!

Breathe. No one is insulting your intelligence nor disputing your point. Also, you've only posted in this thread once so "multiple times" is inaccurate. It's not like anyone is ignoring you.

Now, my benchmarks were just as kind of a general reference as to how much of a difference there is as a general rule, not specifically measuring bus speeds or anything else. I've edited my last post to better differentiate that my benchmarks were just an experiment I was running. It's not necessarily directly in support of IT Sage.

@Sky11 that's what Dell had lead us to believe with the Alienware systems as well, but that perhaps might not be the case. Maybe we can help IT Sage out so we can maybe more accurately speak to that point.

2018 Mac mini: Core i3, 8GB RAM, Sonnet 350 (aftermarket 550w PSU)+XFX Radeon RX-480 8GB Black Edition
2012 Mac mini: Core i7, 16GB RAM, Toshiba 1TB SSD, Seagate 1TB HDD
Custom: Ryzen 7 1700, 16GB Corasir DDR4-3200MHz RAM, 2x Corsair 500GB Neutron SSD, Seagate 3TB SSD, EVGA Nvidia 980Ti
YT:


ReplyQuote
itsage
(@itsage)
Illustrious Member Admin
Joined: 4 years ago
 

I made a post on reddit/r/alienware to ask for more samples. Hopefully we can get more information.

external graphics card builds
best laptops for external GPU
eGPU enclosure buyer's guide

 
2020 13" MacBook Pro [10th,4C,G] + RTX 2080 Ti @ 32Gbps-TB3 (AORUS Gaming Box) + Win10 2004 [build link]  


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 3 years ago
 

You can do that, but need to match cpu when comparing, like always.

If we can thrust the Linus TechTips video, they compared both implementation, but using the same cpu computing raw power.

And already the difference was not clearly visible in their test, only when the TB3 were pushed to his limit, to follow up the cpu/gpu computation.

I found more interesting the last result you posed, were the AGA internal push 10fps more against the core internal on Tom Clancy, showing the importance of the 4x pci-e against the 2x TB, but on the external there is no difference.

The TB3 perform well at the end, but if you plan to really compare the two, you would go after checking the minimum fps and drops. That's the only thing that Linus video pointed out. In situations where the graphic load is intense, the pci-e AGA show higher minimum fps and less frame drop. But the overall average fps stay the same.

2012 13-inch Dell Latitude E6320 + R9 [email protected] (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo Thinkpad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


ReplyQuote
jim_survak
(@jim_survak)
Eminent Member
Joined: 3 years ago
 

@itsage Cool, hopefully we get some info. When I get home I'll check into what you asked. Since my Alienware 13 doesn't have a Thunderbolt port I can't look at that for you. Is there anything else I might be able to help with?

And yes, wimpzilla, I'll put the GPU under some stress to see if CUDA-Z reports differently. 

2018 Mac mini: Core i3, 8GB RAM, Sonnet 350 (aftermarket 550w PSU)+XFX Radeon RX-480 8GB Black Edition
2012 Mac mini: Core i7, 16GB RAM, Toshiba 1TB SSD, Seagate 1TB HDD
Custom: Ryzen 7 1700, 16GB Corasir DDR4-3200MHz RAM, 2x Corsair 500GB Neutron SSD, Seagate 3TB SSD, EVGA Nvidia 980Ti
YT:


ReplyQuote
wimpzilla
(@wimpzilla)
Honorable Member
Joined: 3 years ago
 
Posted by: jim_survak

@itsage Cool, hopefully we get some info. When I get home I'll check into what you asked. Since my Alienware 13 doesn't have a Thunderbolt port I can't look at that for you. Is there anything else I might be able to help with?

And yes, wimpzilla, I'll put the GPU under some stress to see if CUDA-Z reports differently. 

The thing you can do if you have a motherboard with overclocking capabilities:

  1. Disable the Multi-thread option to transform the 3770K into a true quad only and or disable 2 cores to match the i54210U 2 cores 4 thread.
  2. Downclock your 3770K to the same IPC performance of your i5-4210U, check the IPC/multithread performances with cpu-z/cinebench.
  3. Run again the same test and check if you get the same overall performances.
  4. At this point report the bench result and more important your personal gaming feeling experience.

2012 13-inch Dell Latitude E6320 + R9 [email protected] (EXP GDC 8.4) + Win10
E=Mc²

 
2012 15" Lenovo Thinkpad T530 [2nd,4C,Q] + R9 270X @ 4Gbps-mPCIe2 (EXP GDC 8.4) + Win10 [build link]  


ReplyQuote
jim_survak
(@jim_survak)
Eminent Member
Joined: 3 years ago
 

Well... that will prove to be difficult since I don't have that setup anymore. My roommate has a PC I built with an i5-6500 (I think that's what was in it). I can try to disable a couple cores there & underclock it. Though that board does not have a PCIe x4 slot. Or I can use my Ryzen system, which has a PCIe x4 slot. But that's a very different CPU & chipset so the PCIe might also be a very different & therefore give a very different result.

2018 Mac mini: Core i3, 8GB RAM, Sonnet 350 (aftermarket 550w PSU)+XFX Radeon RX-480 8GB Black Edition
2012 Mac mini: Core i7, 16GB RAM, Toshiba 1TB SSD, Seagate 1TB HDD
Custom: Ryzen 7 1700, 16GB Corasir DDR4-3200MHz RAM, 2x Corsair 500GB Neutron SSD, Seagate 3TB SSD, EVGA Nvidia 980Ti
YT:


ReplyQuote
 of  4