3dmark Time Spy DX12 test

Shintai said:
Seems like you are making up your own definition of async.
Click to expand...

No he's not making it up. It's called async shaders, and I've been telling him async shaders isn't synonymous with async compute for at least four months

time-spy-processing-per-frame-100671642-orig.png


It's also made eminently clear that context switching will still be expensive;
OYXqrNE.png
HkdoQxu.png
 
Last edited:
Ieldra said:
No he's not making it up. It's called async shaders, and I've been telling him async shaders isn't synonymous with async compute for at least four months
Click to expand...
You haven't been telling me anything for 4 months. Async as the term is recognized by most of the community mean the concurrent execution of compute and graphics tasks/ques when in discussion of DX12. Posters like you and the rest of the Nvidia supporters take liberties with the term just to make the topic mirky enough to seem plausible like saying async compute which only refers to compute and not graphics. Maxwell could do neither concurrent execution nor dispatch(think I will use this word seeing you guys try far too hard to misconstrue the others). AMD can easily do both. For now it seems Pascal can do concurrent execution but not dispatch (although I am unsure why looking at their die-map). Also what is still up in the air is the ability to communicate with a number of cores simultaneously, which we know AMD is able to do as they have illustrated their hardware's ability to do so for years now.

I have tried to keep it simple and be clear with my terminology, something you are not doing just to make your plight seem correct when there are enough others posting theories otherwise that are reasonably backed up with links and papers.
 
JustReason said:
You haven't been telling me anything for 4 months. Async as the term is recognized by most of the community mean the concurrent execution of compute and graphics tasks/ques when in discussion of DX12. Posters like you and the rest of the Nvidia supporters take liberties with the term just to make the topic mirky enough to seem plausible like saying async compute which only refers to compute and not graphics.
Click to expand...

What ? " Posters like you and the rest of the Nvidia supporters take liberties with the term just to make the topic mirky enough to seem plausible like saying async compute which only refers to compute and not graphics."

QUOTE ME.

I never said anything even remotely like that, that doesn't even make sense, it sounds like something you would say.

I don't give a shit what the "community" recognizes, particularly because I don't know which "community" you're referring to.

Concurrent execution of graphics + compute is exactly what it sounds like, and there is absolutely no requirement for this to happen per CU. That's it. Maxwell can do concurrent execution; on different units, partitioned statically. Parallel execution in fact...

You're not even making sense mate.

Please, show me your theories backed up by evidence. You don't post evidence for shit, ever. You can't even interpret what's written in the whitepaper as I've PROVED in my previous posts. Quite frankly, I don't think you understand it, and I think you're just parroting whatever you find written by some perceived (by you) authority.

Oh, and congratulations on using the word "dispatch", don't know why you made a big deal out of it. Now you're over the glee of having used the word, explain what you mean. Pascal cannot do concurrent dispatch? What does that mean ?

Also what you call a "die-map" is a just an inaccurate, high level representation of the GPU logic and structure, not that I understand what problem you have with it.

I absolutely have been telling you all of this for several months, mark my words mate, when I say I said something, unlike you, I know I'm going to be able to go back and quote myself and link to thread. And also to link to numerous posts made by you in which you spout incoherent gobbledygook
 
Last edited:
Ieldra said:
You're not even making sense mate.
Click to expand...

I think it's intentional and we're being trolled. Hard.

The ignoring of facts, random speculation and constant throwing together of inconsistent claims can't be for real.
 
MangoSeed said:
I think it's intentional and we're being trolled. Hard.

The ignoring of facts, random speculation and constant throwing together of inconsistent claims can't be for real.
Click to expand...

This is very low quality trolling, I think this is genuine misunderstanding + obstinacy
 
MangoSeed said:
I think it's intentional and we're being trolled. Hard.

The ignoring of facts, random speculation and constant throwing together of inconsistent claims can't be for real.
Click to expand...
no trolling just discussing what I read and what I gathered from Nvidias whitepaper. Besides if I am wrong somewhere prove it. Give me your take. It's just a discussion.
 
Im just going to leave this here. Nvidia, DX12, async in the same sentence? yeah right....
3DMark-Time-Spy-DirectX-12-Benchmark_5.jpg
 
JustReason said:
no trolling just discussing what I read and what I gathered from Nvidias whitepaper. Besides if I am wrong somewhere prove it. Give me your take. It's just a discussion.
Click to expand...

What you gathered from the whitepaper is wrong
 
wantapple said:
Im just going to leave this here. Nvidia, DX12, async in the same sentence? yeah right....
Click to expand...

So how are the Pascal cards gaining performance by using it?

Also what does this picture prove exactly?
 
Ieldra said:
So how are the Pascal cards gaining performance by using it?

Also what does this picture prove exactly?
Click to expand...

that was 7 months ago when it was under development... i will trust more real games than this benchmark.
That Nvidia logo makes it completely untrustworthy and biased, wont base any descisions on that test.


 
JustReason said:
Besides if I am wrong somewhere prove it. Give me your take. It's just a discussion.
Click to expand...

These are the facts. Everything else is speculation.

1) nVidia claims that Maxwell can support concurrent execution of graphics + compute via static partitioning of SMs. This is not a great solution for obvious reasons.

2) nVidia claims that Pascal can support concurrent execution of graphics + compute via dynamic hardware load balancing. How this is actually achieved has not been disclosed by nVidia.

3) A test written by an independent developer demonstrates that Pascal sees gains from async execution (see B3D thread I linked earlier).

4) That same test indicates that the high priority compute queue on Maxwell waits for the current graphics draw call to complete. On Pascal it context switches almost immediately.
 
wantapple said:
that was 7 months ago when it was under development... i will trust more real games than this benchmark.
Click to expand...

So ?

oh well, you do that, trust whatever you want man :p Must be hard being so picky with your benchmarks

MangoSeed said:
These are the facts. Everything else is speculation.

1) nVidia claims that Maxwell can support concurrent execution of graphics + compute via static partitioning of SMs. This is not a great solution for obvious reasons.

2) nVidia claims that Pascal can support concurrent execution of graphics + compute via dynamic hardware load balancing. How this is actually achieved has not been disclosed by nVidia.

3) A test written by an independent developer demonstrates that Pascal sees gains from async execution (see B3D thread I linked earlier).

4) That same test indicates that the high priority compute queue on Maxwell waits for the current graphics draw call to complete. On Pascal it context switches almost immediately.
Click to expand...


Can you link to the relevant posts in the thread? You linked to a very early page. Where are the Pascal benches ?
 
MangoSeed said:
Post #1434

DX12 Performance Discussion And Analysis Thread
Click to expand...

Yeah I ran it and logged with GPUView.

Compute only:
1. 1.13ms
2. 1.15ms
3. 1.17ms
4. 1.14ms
5. 1.15ms
6. 1.15ms
7. 1.17ms
8. 1.15ms
9. 1.15ms
10. 1.15ms
11. 1.15ms
12. 1.17ms
13. 1.15ms
14. 1.15ms
15. 1.15ms
16. 1.15ms
17. 1.17ms
18. 1.14ms
19. 1.15ms
20. 1.15ms
21. 1.15ms
22. 1.17ms
23. 1.14ms
24. 1.15ms
25. 1.16ms
26. 1.15ms
27. 1.18ms
28. 1.15ms
29. 1.15ms
30. 1.16ms
31. 1.16ms
32. 1.16ms
33. 2.26ms
34. 2.28ms
35. 2.27ms
36. 2.27ms
37. 2.25ms
38. 2.34ms
39. 2.28ms
40. 2.37ms
41. 2.28ms
42. 2.28ms
43. 2.28ms
44. 2.26ms
45. 2.29ms
46. 2.29ms
47. 2.39ms
48. 2.28ms
49. 2.29ms
50. 2.29ms
51. 2.28ms
52. 2.30ms
53. 2.30ms
54. 2.30ms
55. 2.30ms
56. 2.29ms
57. 2.30ms
58. 2.30ms
59. 2.27ms
60. 2.34ms
61. 2.36ms
62. 2.40ms
63. 2.29ms
64. 2.29ms
65. 3.41ms
66. 3.43ms
67. 3.42ms
68. 3.41ms
69. 3.44ms
70. 3.42ms
71. 3.46ms
72. 3.43ms
73. 3.42ms
74. 3.41ms
75. 3.43ms
76. 3.42ms
77. 3.42ms
78. 3.42ms
79. 3.41ms
80. 3.47ms
81. 3.42ms
82. 3.46ms
83. 3.43ms
84. 3.43ms
85. 3.41ms
86. 3.44ms
87. 3.42ms
88. 3.45ms
89. 3.44ms
90. 3.43ms
91. 3.43ms
92. 3.44ms
93. 3.43ms
94. 3.47ms
95. 3.44ms
96. 3.43ms
97. 4.52ms
98. 4.53ms
99. 4.52ms
100. 4.54ms
101. 4.53ms
102. 4.52ms
103. 4.59ms
104. 4.53ms
105. 4.56ms
106. 4.53ms
107. 4.52ms
108. 4.65ms
109. 4.57ms
110. 4.54ms
111. 4.54ms
112. 4.53ms
113. 4.60ms
114. 4.61ms
115. 4.54ms
116. 4.56ms
117. 4.55ms
118. 4.55ms
119. 4.56ms
120. 4.54ms
121. 4.55ms
122. 4.61ms
123. 4.55ms
124. 4.55ms
125. 4.55ms
126. 4.55ms
127. 4.54ms
128. 4.56ms
Graphics only: 19.48ms (110.24G pixels/s)
Graphics + compute:
1. 20.67ms (103.90G pixels/s)
2. 20.68ms (103.84G pixels/s)
3. 20.79ms (103.28G pixels/s)
4. 20.66ms (103.92G pixels/s)
5. 20.65ms (103.97G pixels/s)
6. 20.77ms (103.38G pixels/s)
7. 20.64ms (104.04G pixels/s)
8. 20.68ms (103.85G pixels/s)
9. 20.70ms (103.74G pixels/s)
10. 20.74ms (103.54G pixels/s)
11. 20.69ms (103.81G pixels/s)
12. 20.71ms (103.68G pixels/s)
13. 20.64ms (104.03G pixels/s)
14. 20.73ms (103.60G pixels/s)
15. 20.73ms (103.57G pixels/s)
16. 20.65ms (103.98G pixels/s)
17. 20.69ms (103.78G pixels/s)
18. 20.71ms (103.68G pixels/s)
19. 20.67ms (103.88G pixels/s)
20. 20.77ms (103.39G pixels/s)
21. 20.76ms (103.42G pixels/s)
22. 20.65ms (103.99G pixels/s)
23. 20.67ms (103.87G pixels/s)
24. 20.76ms (103.46G pixels/s)
25. 20.64ms (104.05G pixels/s)
26. 20.69ms (103.79G pixels/s)
27. 20.75ms (103.50G pixels/s)
28. 20.65ms (103.98G pixels/s)
29. 20.65ms (103.98G pixels/s)
30. 20.74ms (103.55G pixels/s)
31. 20.69ms (103.79G pixels/s)
32. 20.66ms (103.93G pixels/s)
33. 21.92ms (97.99G pixels/s)
34. 21.76ms (98.68G pixels/s)
35. 21.83ms (98.39G pixels/s)
36. 21.83ms (98.38G pixels/s)
37. 21.85ms (98.30G pixels/s)
38. 21.86ms (98.26G pixels/s)
39. 21.92ms (97.96G pixels/s)
40. 21.78ms (98.61G pixels/s)
41. 21.88ms (98.15G pixels/s)
42. 21.79ms (98.57G pixels/s)
43. 21.90ms (98.06G pixels/s)
44. 21.79ms (98.57G pixels/s)
45. 21.83ms (98.38G pixels/s)
46. 21.80ms (98.50G pixels/s)
47. 21.89ms (98.09G pixels/s)
48. 21.82ms (98.43G pixels/s)
49. 21.97ms (97.74G pixels/s)
50. 21.88ms (98.14G pixels/s)
51. 21.93ms (97.91G pixels/s)
52. 21.86ms (98.26G pixels/s)
53. 21.88ms (98.15G pixels/s)
54. 21.88ms (98.15G pixels/s)
55. 21.84ms (98.32G pixels/s)
56. 21.89ms (98.10G pixels/s)
57. 21.86ms (98.25G pixels/s)
58. 21.86ms (98.22G pixels/s)
59. 21.94ms (97.87G pixels/s)
60. 21.86ms (98.25G pixels/s)
61. 21.88ms (98.13G pixels/s)
62. 21.83ms (98.39G pixels/s)
63. 21.80ms (98.51G pixels/s)
64. 21.86ms (98.25G pixels/s)
65. 22.92ms (93.71G pixels/s)
66. 23.11ms (92.93G pixels/s)
67. 23.04ms (93.22G pixels/s)
68. 22.95ms (93.57G pixels/s)
69. 22.95ms (93.56G pixels/s)
70. 22.98ms (93.44G pixels/s)
71. 23.05ms (93.16G pixels/s)
72. 22.94ms (93.61G pixels/s)
73. 23.10ms (92.96G pixels/s)
74. 23.04ms (93.21G pixels/s)
75. 22.91ms (93.72G pixels/s)
76. 22.97ms (93.47G pixels/s)
77. 22.93ms (93.67G pixels/s)
78. 23.02ms (93.29G pixels/s)
79. 22.98ms (93.43G pixels/s)
80. 22.95ms (93.58G pixels/s)
81. 23.12ms (92.89G pixels/s)
82. 23.10ms (92.95G pixels/s)
83. 22.97ms (93.48G pixels/s)
84. 22.95ms (93.55G pixels/s)
85. 22.89ms (93.81G pixels/s)
86. 23.02ms (93.30G pixels/s)
87. 23.10ms (92.96G pixels/s)
88. 22.91ms (93.72G pixels/s)
89. 23.10ms (92.96G pixels/s)
90. 22.97ms (93.49G pixels/s)
91. 23.01ms (93.35G pixels/s)
92. 22.99ms (93.42G pixels/s)
93. 23.01ms (93.34G pixels/s)
94. 23.07ms (93.07G pixels/s)
95. 22.92ms (93.70G pixels/s)
96. 22.93ms (93.67G pixels/s)
97. 24.27ms (88.48G pixels/s)
98. 24.13ms (89.01G pixels/s)
99. 24.13ms (89.01G pixels/s)
100. 24.16ms (88.89G pixels/s)
101. 24.04ms (89.32G pixels/s)
102. 24.14ms (88.98G pixels/s)
103. 24.17ms (88.84G pixels/s)
104. 24.18ms (88.80G pixels/s)
105. 24.23ms (88.61G pixels/s)
106. 24.08ms (89.17G pixels/s)
107. 24.12ms (89.04G pixels/s)
108. 24.21ms (88.70G pixels/s)
109. 24.06ms (89.24G pixels/s)
110. 24.16ms (88.88G pixels/s)
111. 24.19ms (88.77G pixels/s)
112. 24.13ms (88.98G pixels/s)
113. 24.11ms (89.06G pixels/s)
114. 24.09ms (89.14G pixels/s)
115. 24.20ms (88.72G pixels/s)
116. 24.12ms (89.02G pixels/s)
117. 24.17ms (88.85G pixels/s)
118. 24.12ms (89.02G pixels/s)
119. 24.21ms (88.72G pixels/s)
120. 24.26ms (88.53G pixels/s)
121. 24.13ms (89.00G pixels/s)
122. 24.05ms (89.30G pixels/s)
123. 24.07ms (89.22G pixels/s)
124. 24.12ms (89.02G pixels/s)
125. 24.16ms (88.89G pixels/s)
126. 24.14ms (88.98G pixels/s)
127. 24.09ms (89.16G pixels/s)
128. 24.21ms (88.70G pixels/s)
Graphics, compute single commandlist:
1. 20.73ms (103.58G pixels/s)
2. 20.66ms (103.96G pixels/s)
3. 20.69ms (103.78G pixels/s)
4. 20.65ms (104.01G pixels/s)
5. 20.63ms (104.12G pixels/s)
6. 20.69ms (103.82G pixels/s)
7. 20.73ms (103.61G pixels/s)
8. 20.70ms (103.75G pixels/s)
9. 20.67ms (103.91G pixels/s)
10. 20.66ms (103.97G pixels/s)
11. 20.65ms (103.98G pixels/s)
12. 20.67ms (103.87G pixels/s)
13. 20.75ms (103.51G pixels/s)
14. 20.65ms (104.00G pixels/s)
15. 20.72ms (103.65G pixels/s)
16. 20.66ms (103.96G pixels/s)
17. 20.61ms (104.19G pixels/s)
18. 20.74ms (103.54G pixels/s)
19. 20.70ms (103.75G pixels/s)
20. 20.67ms (103.89G pixels/s)
21. 20.73ms (103.60G pixels/s)
22. 20.64ms (104.03G pixels/s)
23. 20.58ms (104.35G pixels/s)
24. 20.65ms (103.97G pixels/s)
25. 20.69ms (103.82G pixels/s)
26. 20.68ms (103.84G pixels/s)
27. 20.74ms (103.55G pixels/s)
28. 20.64ms (104.05G pixels/s)
29. 20.67ms (103.91G pixels/s)
30. 20.72ms (103.65G pixels/s)
31. 20.72ms (103.64G pixels/s)
32. 20.72ms (103.66G pixels/s)
33. 21.79ms (98.55G pixels/s)
34. 21.76ms (98.71G pixels/s)
35. 21.76ms (98.68G pixels/s)
36. 21.74ms (98.79G pixels/s)
37. 21.79ms (98.56G pixels/s)
38. 21.83ms (98.38G pixels/s)
39. 21.83ms (98.36G pixels/s)
40. 21.74ms (98.79G pixels/s)
41. 21.81ms (98.44G pixels/s)
42. 21.80ms (98.52G pixels/s)
43. 21.75ms (98.74G pixels/s)
44. 21.88ms (98.14G pixels/s)
45. 21.78ms (98.60G pixels/s)
46. 21.77ms (98.64G pixels/s)
47. 21.86ms (98.25G pixels/s)
48. 21.84ms (98.33G pixels/s)
49. 21.79ms (98.57G pixels/s)
50. 21.83ms (98.37G pixels/s)
51. 21.77ms (98.64G pixels/s)
52. 21.86ms (98.23G pixels/s)
53. 21.79ms (98.56G pixels/s)
54. 21.86ms (98.26G pixels/s)
55. 21.85ms (98.27G pixels/s)
56. 21.77ms (98.65G pixels/s)
57. 21.74ms (98.76G pixels/s)
58. 21.85ms (98.29G pixels/s)
59. 21.78ms (98.61G pixels/s)
60. 21.80ms (98.53G pixels/s)
61. 21.81ms (98.46G pixels/s)
62. 21.77ms (98.64G pixels/s)
63. 21.81ms (98.48G pixels/s)
64. 21.91ms (98.01G pixels/s)
65. 22.92ms (93.68G pixels/s)
66. 22.85ms (93.97G pixels/s)
67. 22.98ms (93.43G pixels/s)
68. 22.86ms (93.93G pixels/s)
69. 22.93ms (93.65G pixels/s)
70. 22.86ms (93.93G pixels/s)
71. 22.91ms (93.75G pixels/s)
72. 23.04ms (93.21G pixels/s)
73. 22.89ms (93.83G pixels/s)
74. 22.92ms (93.71G pixels/s)
75. 22.95ms (93.56G pixels/s)
76. 22.96ms (93.51G pixels/s)
77. 22.90ms (93.80G pixels/s)
78. 22.90ms (93.77G pixels/s)
79. 22.89ms (93.80G pixels/s)
80. 22.96ms (93.54G pixels/s)
81. 22.92ms (93.69G pixels/s)
82. 22.89ms (93.81G pixels/s)
83. 23.03ms (93.23G pixels/s)
84. 22.87ms (93.91G pixels/s)
85. 22.97ms (93.48G pixels/s)
86. 22.88ms (93.86G pixels/s)
87. 22.97ms (93.50G pixels/s)
88. 22.93ms (93.65G pixels/s)
89. 22.88ms (93.85G pixels/s)
90. 22.87ms (93.90G pixels/s)
91. 22.94ms (93.60G pixels/s)
92. 22.95ms (93.57G pixels/s)
93. 22.87ms (93.88G pixels/s)
94. 22.99ms (93.41G pixels/s)
95. 22.89ms (93.82G pixels/s)
96. 22.93ms (93.65G pixels/s)
97. 24.00ms (89.48G pixels/s)
98. 24.01ms (89.44G pixels/s)
99. 24.00ms (89.46G pixels/s)
100. 23.96ms (89.61G pixels/s)
101. 23.98ms (89.55G pixels/s)
102. 23.97ms (89.59G pixels/s)
103. 24.02ms (89.42G pixels/s)
104. 24.09ms (89.15G pixels/s)
105. 24.00ms (89.47G pixels/s)
106. 24.01ms (89.43G pixels/s)
107. 24.00ms (89.48G pixels/s)
108. 24.10ms (89.11G pixels/s)
109. 24.02ms (89.41G pixels/s)
110. 24.05ms (89.28G pixels/s)
111. 24.03ms (89.35G pixels/s)
112. 24.19ms (88.79G pixels/s)
113. 24.04ms (89.33G pixels/s)
114. 24.08ms (89.17G pixels/s)
115. 24.01ms (89.45G pixels/s)
116. 24.01ms (89.44G pixels/s)
117. 24.11ms (89.05G pixels/s)
118. 24.08ms (89.17G pixels/s)
119. 24.02ms (89.39G pixels/s)
120. 24.06ms (89.25G pixels/s)
121. 24.00ms (89.47G pixels/s)
122. 24.06ms (89.25G pixels/s)
123. 24.03ms (89.37G pixels/s)
124. 23.98ms (89.54G pixels/s)
125. 24.09ms (89.14G pixels/s)
126. 24.00ms (89.47G pixels/s)
127. 24.03ms (89.38G pixels/s)
128. 24.01ms (89.44G pixels/s)
Latency (compute starts 10ms after graphics):
1. 10.40ms
2. 10.83ms
3. 10.26ms
4. 10.29ms
5. 10.26ms
6. 10.20ms
7. 11.20ms
8. 10.19ms
9. 10.19ms
10. 10.19ms
11. 11.20ms
12. 10.19ms
13. 10.19ms
14. 10.19ms
15. 10.20ms
16. 10.20ms
17. 10.76ms
18. 10.19ms
19. 10.19ms
20. 10.71ms
21. 11.20ms
22. 10.19ms
23. 10.19ms
24. 10.20ms
25. 10.19ms
26. 10.21ms
27. 10.64ms
28. 10.66ms
29. 10.70ms
30. 10.37ms
31. 10.19ms
32. 10.75ms
33. 10.59ms
34. 10.70ms
35. 10.20ms
36. 10.68ms
37. 10.19ms
38. 10.70ms
39. 10.67ms
40. 10.19ms
41. 10.19ms
42. 10.67ms
43. 11.20ms
44. 10.19ms
45. 11.19ms
46. 10.19ms
47. 10.20ms
48. 10.23ms
49. 10.19ms
50. 10.19ms
51. 10.19ms
52. 10.78ms
53. 10.19ms
54. 10.19ms
55. 10.19ms
56. 10.76ms
57. 10.70ms
58. 11.19ms
59. 10.19ms
60. 11.19ms
61. 10.70ms
62. 10.29ms
63. 10.19ms
64. 10.74ms
65. 10.19ms
66. 10.25ms
67. 11.20ms
68. 10.19ms
69. 10.19ms
70. 10.70ms
71. 10.42ms
72. 10.22ms
73. 10.19ms
74. 10.70ms
75. 10.24ms
76. 10.19ms
77. 10.75ms
78. 11.19ms
79. 10.20ms
80. 10.20ms
81. 10.68ms
82. 10.19ms
83. 10.19ms
84. 11.66ms
85. 10.61ms
86. 10.30ms
87. 10.19ms
88. 10.28ms
89. 10.19ms
90. 10.55ms
91. 10.19ms
92. 10.65ms
93. 10.70ms
94. 10.19ms
95. 10.66ms
96. 10.70ms
97. 10.19ms
98. 10.69ms
99. 10.19ms
100. 10.19ms
101. 10.19ms
102. 10.71ms
103. 10.70ms
104. 10.19ms
105. 10.19ms
106. 11.20ms
107. 10.40ms
108. 10.19ms
109. 10.33ms
110. 10.71ms
111. 10.20ms
112. 10.19ms
113. 10.19ms
114. 10.19ms
115. 10.19ms
116. 11.63ms
117. 10.19ms
118. 9.75ms
119. 10.19ms
120. 10.20ms
121. 10.17ms
122. 10.19ms
123. 10.19ms
124. 11.20ms
125. 10.70ms
126. 10.19ms
127. 11.20ms
128. 10.19ms
Average: 10.45ms

CxUWWVT.png
 
Last edited:
wantapple said:
That Nvidia logo makes it completely untrustworthy and biased, wont base any descisions on that test.


Click to expand...

To be able to live in such a paranoid and delusional world. You do realize 3Dmark would have a hell of a lot to lose if it somehow was GPU biased?
 
About to test this out but for some odd reason 3dMark never updates within the App. I always need to download the damn thing again.
 
yay are we getting back on topic?! i'd like to see results not bickering for a page and a half...
 
  • Like
Reactions: N4CR
like this
Presbytier said:
To be able to live in such a paranoid and delusional world. You do realize 3Dmark would have a hell of a lot to lose if it somehow was GPU biased?
Click to expand...

wantapple is fine with using AotS in in situation where they beat Nvidia, even though that is an AMD sponsored game:
Doom Alpha GPU test
Ashes of the Singularity

If people start throwing out all of the games used in marketing by one company or another, there won't be anything left to benchmark :p
 
I just got 7234.
7425 for Graphics and 6316 for CPU test with Ncase in my sig.
 
Performance is great. I'm gaming at 1440P, but do triple monitors and I'm able to get smooth gameplay at 7680x1440 (so a bit larger than 4K).

The Nvidia driver is still a little bugged with everything enabled (meaning Surround, G-Sync, and SLI combined) but with one monitor it's pretty solid.
 
All this shit about nvidia this amd this bla bla bla bla.

facts!
  • Amd gains much more from Async (10-20%) as more shaders are utiliized
  • Maxwell loses performance.
  • Pascal gains like 2-3%, sometimes nothing.
AMD wins in Async. Nvidia has pure performance. (Its obvious as more titles come out that utilize dx12 and vulkan amd will be able to close the gap. good for them)

For the sake of argument its obvious amd is better at this certain feature than nvidia. What is there to argue, really why one is butt hurt over it if you have nvidia card is out of my grasp. Nvidia is faster hands down so why this need to justify that some how it doesn't need async. Good for people that have GCN cards because this just gives them more life. What's to hate here, just cuz you got one card over the other doesn't mean obvious isn't obvious.
 
wantapple said:
Im just going to leave this here. Nvidia, DX12, async in the same sentence? yeah right....
3DMark-Time-Spy-DirectX-12-Benchmark_5.jpg
Click to expand...
So that was probably NVIDIA sponsored event... I bet you think that Ashes of the Singularity is a fine benchmark despite AMD logos in every event and even on their site. You'd just happily trust that?

Sometimes I just don't get......
 
You must log in or register to reply here.
Back
Top