The Landscape of Parallel Computing Research: A View from Berkeley
2006 · paper · Computer science
Reports a Berkeley group's view of the shift to parallel microprocessors, arguing that simply adding cores will give diminishing returns. It recommends targeting thousands of cores, evaluating systems with 13 "dwarfs" of computation, and using autotuners and more human-centric programming models.
Checked: PDF from www2.eecs.berkeley.edu, fetched 4 Oct 2026. Page numbers are those of that PDF.
11 findings
Substance Section 4.1.1 says "superscalar features adversely affected the energy-delay product. The power overhead needed for additional hardware did not outweigh the performance benefits." The second sentence says the opposite of the first and of the sentence that follows it.
- p. 24 Section 4.1.1, Gonzalez and Horowitz paragraph
The power overhead needed for additional hardware did not outweigh the performance benefits.
Only the authors can say what is right here.
Judge's note: Seen on the p.24 crop. If superscalar features worsen the energy-delay product, and give 'modest increases in performance at the cost of significant area and energy overhead', then the overhead did outweigh the benefits. The sentence appears to have its terms reversed. Only the authors can confirm the intended wording.
Second check: another model agreed
Report a problem with this finding
Substance Section 4.1.3 prints "SpeedupHomogeneous = 1 / (0.1 – 0.9/100) = 9.2" and "SpeedupHeterogeneous = 1 / (0.1/2 – 0.9/90) = 16.7"; under Amdahl's Law the terms add, and the printed results follow only with '+'.
- p. 25 Section 4.1.3, speedup equations
SpeedupHomogeneous = 1 / (0.1 – 0.9/100) = 9.2 times faster
- p. 25 Section 4.1.3, speedup equations
SpeedupHeterogeneous = 1 / (0.1/2 – 0.9/90) = 16.7 times faster
Judge's note: The p.25 crop clearly shows an en dash (minus) in both formulas. As printed, the expressions give 10.99 and 25. With '+' they give 9.17 and 16.67, which match the printed values.
Second check: another model agreed
Report a problem with this finding
Reference or label Section 4.2 says "as mentioned in CW #8 in Section 2, the number of processor cycles to access main memory has grown"; the Memory wall is CW #7 in Section 2.
- p. 26 Section 4.2, first paragraph
as mentioned in CW #8 in Section 2, the number of processor cycles to
- p. 7 Section 2, CW 7
New CW is the “Memory wall” [Wulf and McKee 1995]
- p. 26 Section 4.2, first paragraph
Alas, as mentioned in CW #8 in Section 2, the number of processor cycles to access main memory has grown dramatically
- p. 26 4.2
as mentioned in CW #8 in Section 2
Judge's note: The p.7 crop shows item 7: 'New CW is the "Memory wall" ... 200 clocks to access DRAM'. The p.26 crop shows 'CW #8'. The pointer is off by one.
Second check: another model agreed
Report a problem with this finding
Reference or label Citation keys do not match reference entries: the text cites "[Woo et al 1992]" (p.9) but the only entry is "[Woo et al 1995]" (p.56); the text cites "[Weinberg 2004]" (p.43) but the entry key is "[Weinburg 2004] B. Weinberg" (p.55).
- p. 9 Section 3.0
[Singh et al 1992] [Woo et al 1992]
- p. 56 References
[Woo et al 1995]
- p. 56 References
[Woo et al 1995] S.C. Woo, M. Ohara, E. Torrie
- p. 43 7.1
[Weinberg 2004]
Judge's note: The p.9 crop shows '[Singh et al 1992] [Woo et al 1992]'. The reference list has only '[Woo et al 1995] S.C. Woo, M. Ohara, E. Torrie, J.P. Singh, and A. Gupta, “The SPLASH-2 Programs...” ISCA ’95'. The p.43 crop shows '[Weinberg 2004]'. The p.55 crop shows the entry key '[Weinburg 2004]' while the author is printed as 'B. Weinberg'. The Vetter and Meuller/Mueller mismatch in reviewer note 3 is judged under C007.
Second check: another model agreed
Report a problem with this finding
Reference or label "[Russell 1976]" (p.10) and "[Greengard and Rokhlin 1987]" (Figure 3, p.12) are cited but have no entry in the reference list.
- p. 10 Section 3.1
vector architectures for linear algebra [Russell 1976]
- p. 12 Figure 3, N-body row
Method [Greengard and Rokhlin 1987]
- p. 12 Figure 3, Structured Grids row
[LBMHD 2005]
Judge's note: I saw both citations on the page images. A search of the full text finds 'Russell' and 'Greengard' only at the citation sites (lines 368 and 523), never in the reference list. The list runs Rowen and Leibson → Scott and Granlund → Gries, with no Russell or Greengard entry. The missing [LBMHD 2005] entry is judged under C012.
Second check: another model agreed
Report a problem with this finding
Reference or label Figure 3 cites "[LBMHD 2005]", but no reference has that key. The lattice-Boltzmann entry "[Vahala et al 2005] G. Vahala..." is never cited.
- p. 12 Figure 3, Structured Grids row
[LBMHD 2005]
- p. 55 References
[Vahala et al 2005] G. Vahala
Only the authors can say what is right here.
Judge's note: The p.12 crop shows 'Magneto-hydrodynamics [LBMHD 2005])'. A search of the full text finds 'LBMHD' only at that citation, and 'Vahala' only in its own reference entry (line 3058). The citation probably points to Vahala et al 2005, but the keys do not match.
Second check: another model agreed
Report a problem with this finding
Reference or label Section 3.4 says "examination of machine learning, databases, and graphics in Sections 3.2.1 to 3.3.3"; there is no Section 3.3.3, and the intended range is 3.2.1 to 3.2.3.
- p. 17 Section 3.4, last sentence
machine learning, databases, and graphics in Sections 3.2.1 to 3.3.3
- p. 17 Section 3.4, end of first paragraph
machine learning, databases, and graphics in Sections 3.2.1 to 3.3.3.
- p. 17 3.4
examination of machine learning, databases, and graphics in Sections 3.2.1 to 3.3.3
- p. 17 3.4
Sections 3.2.1 to 3.3.3
Judge's note: Seen on the p.17 crop. The section headings in the text are 3.2.1 Machine Learning, 3.2.2 Database Software, 3.2.3 Computer Graphics and Games and 3.2.4. Section 3.3 'Composition of Dwarfs' has no subsections.
Second check: another model agreed
Report a problem with this finding
Reference or label "[Amdahl 1967]" is cited in Section 4.1.3, but the reference list has no Amdahl entry.
- p. 25 Section 4.1.3
performance on a parallel computer [Amdahl 1967]
Judge's note: Searched the text: the 'A' references are ABAQUS, Adiletta, Allen, Altschul, Alverson (x2), Arnold, Arvind and Aspuru-Guzik. 'Amdahl 1967' appears only at the in-text citation.
Second check: another model agreed
Report a problem with this finding
Reference or label Section 4.3: "With the exception of the 3D FFT (see Figure 2)" points to the processor-performance plot. The communication patterns are in Figure 3.
- p. 28 Section 4.3, third bullet
With the exception of the 3D FFT (see Figure 2)
- p. 28 4.3
3D FFT (see Figure 2)
Judge's note: Crop C032-a (p.28) shows 'see Figure 2'. The Figure 2 caption (text line 297) reads 'Processor performance improvement between 1978 and 2006 using integer SPEC'. Section 3.1 says 'Figure 3 also shows the inter-processor communication patterns' (line 369). So the pointer should be Figure 3.
Second check: another model agreed
Report a problem with this finding
Reference or label Section 7.1: "performance and energy counters to aid program optimization (see Section 4.5)" points to Dependability. The counters are covered in Section 4.6.
- p. 43 Section 7.1, second paragraph
performance and energy counters to aid program optimization (see Section 4.5)
- p. 44 Section 7.2
counters that can measure these performance issues (see Section 4.6)
- p. 32 heading 4.6
4.6 Performance and Energy Counters
Judge's note: Crop C041-a (p.43) shows 'Section 4.5'. Section 4.5 is 'Dependability' (text line 1715), and crop C041-c (p.32) shows the heading '4.6 Performance and Energy Counters'. Section 7.2 cites 4.6 for the same topic (crop C033-e).
Second check: another model agreed
Report a problem with this finding
Reference or label The [Patterson et al 1997] entry prints the date as "IEEE Micro, vol. 17, no. 2, Mar.–Apr. 1993, pp. 34–44", which conflicts with the 1997 in its key and in the in-text citation.
- p. 53 References
IEEE Micro, vol. 17, no. 2, Mar.–Apr. 1993, pp. 34–44.
- p. 53 References
[Patterson et al 1997] ... IEEE Micro, vol. 17, no. 2, Mar.–Apr. 1993
- p. 53 References
IEEE Micro, vol. 17, no. 2, Mar.–Apr. 1993
- p. 52 References
[Im et al 2005] ... Spr. 2004
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that IEEE Micro vol. 17 is the 1997 volume. Within the paper, the p.53 crop confirms that the key says 1997 while the entry says 1993; the text cites '[Patterson et al 1997]' on p.~29. The Im, Xu and Ebcioglu crops in this cluster repeat C051.
Second check: another model agreed
Report a problem with this finding
30 minor (typos and the like; most checked by the judge only)
Reference or label Reference key "[Xu et al 2003]" conflicts with its own entry: "Proceedings of the 31st Annual International Symposium on Computer Architecture (ISCA ’04), 2004."
- p. 43 Section 7.1
trace collection for debugging [Xu et al 2003]
- p. 56 References
Computer Architecture (ISCA ’04), 2004.
Only the authors can say what is right here.
Judge's note: Crop C043-b (p.56) shows the key dated 2003 and the entry dated 2004 (31st ISCA '04). The text cites it as [Xu et al 2003] (p.43). The files alone cannot show which year is correct, so this goes to the authors.
Second check: another model agreed
Report a problem with this finding
Reference or label In three references the year in the citation key differs from the date printed in the entry: "[Ebcioglu et al 2006] ... (P-PHEC 2005), Feb. 2005", "[Im et al 2005] ... vol. 18, no. 1, Spr. 2004", "[Xu et al 2003] ... (ISCA ’04), 2004".
- p. 50 References
[Ebcioglu et al 2006]
- p. 56 References
[Xu et al 2003]
- p. 52 References
[Im et al 2005]
- p. 56 References
[Xu et al 2003] ... (ISCA ’04), 2004.
Only the authors can say what is right here.
Judge's note: All three entries were checked on the p.50, p.52 and p.56 crops, and each mismatch is as quoted. The in-text citations use the same keys. Nothing in the paper explains the differences (for example a preprint date), so only the authors can say which year is meant.
Second check: another model agreed
Report a problem with this finding
Wording The abstract says the evolutionary approach "may work from 2 or 8 processor systems". 'from 2 or 8' reads like a slip for a range (e.g. 'from 2 to 8').
- p. 3 Abstract
may work from 2 or 8 processor systems
- p. 46 Conclusion
suitable for 2 to 32 processors
Only the authors can say what is right here.
Judge's note: The abstract (p.3) does print 'from 2 or 8 processor systems, but is likely to face diminishing returns as 16 and 32 processor systems are realized'. The conclusion (p.46) says multicore models 'suitable for 2 to 32 processors' cannot evolve to 1000s of processors. That is a different claim (evolving to manycore), not a contradiction, so the cross-section inconsistency is weak. The only defect is the odd 'from 2 or 8' wording, and only the authors know the intended range.
Second check: another model agreed
Report a problem with this finding
Wording Section 3.1 says "SPEC2006 has 29 benchmarks and EEMBC has 41", but Section 3.2 says "from our examination of the 41 EEMBC kernels and the 26 SPEC2006 programs".
- p. 10 3.2
SPEC2006 has 29 benchmarks and EEMBC has 41
- p. 13 3.2
the 41 EEMBC kernels and the 26 SPEC2006 programs
Only the authors can say what is right here.
Judge's note: Both statements are visible on the page images (p.10 and p.13). The EEMBC count (41) agrees between the two, but the SPEC2006 count differs (29 vs 26). No text explains whether 26 is a subset that was examined. Only the authors can say which count is intended.
Second check: another model agreed
Report a problem with this finding
Wording From "188 cores implemented in 130 nm", the paper projects "752 processors in 45nm and 1504 in 30nm". That is a 4x step to 45 nm, whereas doubling per generation over 130→90→65→45 gives 8x (1504) at 45 nm.
- p. 25 Section 4.1.2
752 processors in 45nm and 1504 in 30nm
- p. 24 Section 4.1.2
188 cores implemented in 130 nm technology
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the process-node sequence 130/90/65/45/32 nm. The paper states that core counts 'can double with each generation of silicon' (p.22), but it gives no scaling assumption behind 752 and 1504. Simple area scaling gives about 8.3x at 45 nm. The authors should state the basis for the projection.
Second check: another model agreed
Report a problem with this finding
Typo Acknowledgement on PDF p.2 reads "Wed like to thank the following"; the apostrophe is missing (should be "We'd").
- p. 2 Acknowledgement
Wed like to thank the following
- p. 2
Wed like to thank
Judge's note: The page-2 crop shows 'Wed' with no apostrophe. The same sentence on PDF p.48 (text line 2611) reads "We’d like to thank", so the p.2 form is a slip.
Second check: none (judge only)
Report a problem with this finding
Typo The author line on PDF p.3 prints "Krste Asanovíc": the acute accent sits on the i. Elsewhere (cover, references, p.48) the name is "Asanovic".
- p. 3 Author line
Krste Asanovíc
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the author's name is spelled Asanović (accent on the c). The p.3 image does show 'Asanovíc' in italics. Inside the paper the name appears unaccented as 'Asanovic' (text line 4, and reference entries such as Heo and Asanovic 2004), so the p.3 accent is inconsistent with the rest of the document.
Second check: none (judge only)
Report a problem with this finding
Typo The text cites "[Vetter and Meuller 2002]" (p.10 and again in Section 4.3), but the reference entry is "[Vetter and Mueller 2002] J.S. Vetter and F. Mueller".
- p. 10 Section 3.1
[Vetter and Meuller 2002]
- p. 55 References
[Vetter and Mueller 2002]
Judge's note: The p.10 crop shows 'Meuller'. Text line 1538 repeats '[Vetter and Meuller 2002]'. The reference entry (text line 3066) reads 'Mueller'. The intended work is unambiguous, so this is a misspelled key.
Second check: none (judge only)
Report a problem with this finding
Typo Citation keys do not match: the text cites "[Weinberg 2004]" but the reference is keyed "[Weinburg 2004] B. Weinberg", and the text cites "[Vetter and Meuller 2002]" but the reference is "[Vetter and Mueller 2002] ... F. Mueller".
- p. 43 Section 7.1
[Weinberg 2004]
- p. 55 References
[Weinburg 2004] B. Weinberg
- p. 10 3.1
Vetter and Meuller 2002
- p. 55 References
[Vetter and Mueller 2002]
Judge's note: Crops C042-b (p.55) and C042-a/C033-c (p.43) show the Weinberg/Weinburg mismatch. C042-e (p.10) and C042-d (p.55) show the Meuller/Mueller mismatch. Each entry can still be identified, so this is cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Misspellings in Figure 3: "Block Triadiagonal Matrix" (for Tridiagonal), "DSPs, Zalink PDSP [Zarlink 2006]" (the company is spelled Zarlink in the adjacent citation), and "[Edinburg 2006]".
- p. 11 Figure 3, Dense Linear Algebra row
Triadiagonal
- p. 11 Figure 3, Spectral Methods row
DSPs, Zalink
- p. 12 Figure 3, Structured Grids row
[Edinburg 2006]
- p. 11 Figure 3, Spectral Methods row
DSPs, Zalink PDSP [Zarlink 2006]
Judge's note: I saw all three spellings on the p.11/p.12 crops. 'Zalink' conflicts with '[Zarlink 2006]' in the same cell and with the reference entry. 'Triadiagonal' is not a word. 'Edinburg' also appears in the reference entry 'University of Edinburg'; that it should read Edinburgh rests on outside knowledge of the university's name.
Second check: none (judge only)
Report a problem with this finding
Typo Misspellings: Figure 6 has "princpal", "satisficability" and "SPEC Fl. Pl." (it is "SPEC Fl. Pt." elsewhere in the figure), and its caption has "Graphcs/Games"; Figure 3 has "DSPs, Zalink" (the citation and reference read Zarlink); the text has "drawfs" and "(Sections 5.4)".
- p. 19 Figure 6, Dense Linear Algebra / Machine Learning
princpal
- p. 20 Figure 6, Backtrack row / Machine Learning
satisficability
- p. 21 Figure 6 caption
Graphcs/Games
- p. 19 Figure 6, Dense Linear Algebra / General Purpose
SPEC Fl. Pl.: Hidden Markov
Judge's note: Checked on the upright images of pp.19-21 and the crops for pp.11, 15 and 34. All items are visible. The text layer reads 'Fl. Pl.' only in the dense linear algebra cell; every other cell reads 'Fl. Pt.'. 'drawfs' is also C017, and 'princpal'/'satisficability' are also C022.
Second check: none (judge only)
Report a problem with this finding
Typo Copy-editing slips in Section 3.2: "To go beyond to EEMBC and SPEC", "Its uses a commercial sorting package called Nsort, which does sorts either", "The principle burden for the host processor", "did not extend the drawfs".
- p. 14 Section 3.2.2
Its uses a commercial sorting package called Nsort, which does sorts either
- p. 15 Section 3.2.3
did not extend the drawfs
- p. 15 Section 3.2.3
The principle burden for the host processor
- p. 13 Section 3.2
To go beyond to EEMBC and SPEC
Judge's note: I saw each phrase on the page images (p.13, p.14, p.15). The intended forms are 'go beyond EEMBC', 'It uses', 'which sorts either', 'principal burden' and 'dwarfs'.
Second check: none (judge only)
Report a problem with this finding
Typo Two punctuation errors: "approach. In the era of multicore and manycore. Popular algorithms..." (a full stop leaves a fragment; a comma is needed) and "intermediate key/value pairs, The second phase" (a comma where a full stop belongs).
- p. 16 Section 3.2.4, paragraph after Figure 4
In the era of multicore and manycore. Popular algorithms
- p. 14 3.2.2
intermediate key/value pairs, The second phase
Judge's note: Both are confirmed on the images. In the p.16 crop, 'In the era of multicore and manycore.' stands as a fragment. In the p.14 crop, a comma is followed by a capitalised 'The'.
Second check: none (judge only)
Report a problem with this finding
Typo Section 3.2.3 prints "The principle burden for the host processor"; it should read "principal burden".
- p. 15 Section 3.2.3, second paragraph
The principle burden for the host processor
Judge's note: Seen on the crop of p.15. 'Principle' is the wrong word here; the meaning is 'main'.
Second check: none (judge only)
Report a problem with this finding
Typo Section 3.2.3 prints "did not extend the drawfs beyond the 13 identified above"; it should read "dwarfs".
- p. 15
drawfs beyond the 13 identified above
Judge's note: The page image (p.15 crop) shows 'drawfs', so this is not an extraction artefact. The same item also appears in mixed cluster C021.
Second check: none (judge only)
Report a problem with this finding
Typo Figure 6, Dense Linear Algebra row, Embedded column reads "EEMBC Digital Entertainment: RSA MP3 Decode,"; a comma appears to be missing between RSA and MP3 Decode.
- p. 19 Figure 6, row 1, Embedded Computing column
RSA MP3 Decode
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: that RSA and MP3 Decode are separate EEMBC kernels. Visible on the p.19 upright image. The text treats RSA separately (p.13: 'RSA encryption for data security'), and other rows list 'MP3 Decode' on its own. Listing the same kernel under several dwarfs (rows 1 and 8, etc.) is consistent with the figure's purpose, so that part of the finding is not a defect.
Second check: none (judge only)
Report a problem with this finding
Typo The paper cites "[Harstein and Puzak 2003]" (text and reference list); the author's name is presumably Hartstein.
- p. 24 4.1.1
Harstein and Puzak 2003
Only the authors can say what is right here.
Judge's note: Rests on outside knowledge: the author's surname is spelt Hartstein. The p.24 crop shows 'Harstein', and the reference list entry also reads 'A. Harstein'. The two are consistent within the paper.
Second check: none (judge only)
Report a problem with this finding
Typo Section 4.4.4: "(corresponding to the “graph following” dwarf)". The dwarf is named "Graph Traversal" everywhere else.
- p. 30 Section 4.4.4
(corresponding to the “graph following” dwarf)
- p. 30 4.4.4
corresponding to the “graph following” dwarf
- p. 26 4.2
see Figure 9 in Section 8
- p. 30 4.4.4
“graph following” dwarf
Judge's note: Crop C035-a (p.30) shows 'graph following'. The dwarf lists (text lines 635, 779, 2592) name it 'Graph Traversal'. This is a terminology inconsistency only. The 'Figure 9 in Section 8' pointer (crop C035-c) is fine: Figure 9 sits in the concluding section.
Second check: none (judge only)
Report a problem with this finding
Typo Missing, extra or wrong words: "requirements tend utilize" (p.28), "combined packed switches" (p.29), "is as at least as important" (p.43), "To make help programmers" (p.44), "needed rather be forced" (p.41), "this version will a large-scale example" (p.45).
- p. 28 Section 4.3
requirements tend utilize only a fraction
- p. 29 Section 4.3
design that combined packed switches
- p. 43 Section 7.1
is as at least as important
- p. 45 Section 7.3, RAMP Blue
this version will a large-scale example
Judge's note: Each phrase appears as quoted on crops C033-a, b, c, d, e and f. Evident fixes: 'tend to utilize', 'packet switches', 'is at least as important', 'To help programmers', 'rather than be forced', 'will be a large-scale example'.
Second check: none (judge only)
Report a problem with this finding
Typo Section 4.5: "Virtual Machines can also help systems resilient to failures" is missing a verb ('help make systems resilient').
- p. 32 4.5
can also help systems resilient to failures
Judge's note: Crop C036-a (p.32) shows the phrase as quoted.
Second check: none (judge only)
Report a problem with this finding
Typo Heading "5.4 Models must support of proven styles of parallelism" has a stray 'of', and the Section 5 intro writes "(Sections 5.4)" in the plural for a single section.
- p. 37 5.4 heading
Models must support of proven styles of parallelism
- p. 34 Section 5 intro
(Sections 5.4)
- p. 34
styles of parallelism that have been proven successful in the past (Sections 5.4)
Judge's note: Crops C039-a (p.37) and C039-b (p.34) show both as quoted.
Second check: none (judge only)
Report a problem with this finding
Typo Section 5.1: "formalism-centric.Hardware-centric" has no space after the full stop.
- p. 35 5.1
formalism-centric.Hardware-centric
Judge's note: Crop C037-a (p.35) shows no space in the rendered page, so this is not an extraction artefact.
Second check: none (judge only)
Report a problem with this finding
Typo Section 5.3 list: "64 bits (Integer, Double-precision floating point" and "128 bits (Integer, Quad-Precision floating point" are missing their closing parentheses.
- p. 37 Section 5.3, bullet list
64 bits (Integer, Double-precision floating point
- p. 37 5.3
128 bits (Integer, Quad-Precision floating point
- p. 37
• 64 bits (Integer, Double-precision floating point
Judge's note: Crop C038-a (p.37) shows both bullets without a closing ')'. The other bullets in the list close their parentheses.
Second check: none (judge only)
Report a problem with this finding
Typo Section 6.1.3: "especially with the judicial use of performance models" should read 'judicious'.
- p. 41 6.1.3
the judicial use of performance models
- p. 41
especially with the judicial use of performance models
Judge's note: Crop C040-a (p.41) shows 'judicial'. The context (careful use of models) calls for 'judicious'.
Second check: none (judge only)
Report a problem with this finding
Typo Stray word in Section 7.2: "To make help programmers progress towards these goals" should read "To help programmers progress towards these goals".
- p. 44 Section 7.2
To make help programmers progress towards these goals
Judge's note: Seen on the p.44 crop. The sentence contains the stray word 'make' exactly as quoted.
Second check: none (judge only)
Report a problem with this finding
Typo Missing verb in the RAMP Blue bullet: "Led by the Berkeley contingent, this version will a large-scale example using MPI" should read "will be a large-scale example".
- p. 45 7.3
this version will a large-scale example
- p. 45
this version will a large-scale example using MPI
Judge's note: The p.45 crop shows 'this version will' at the end of one line and 'a large-scale example using MPI' at the start of the next. No verb is printed between them.
Second check: none (judge only)
Report a problem with this finding
Typo The Figure 9 caption reads "inspired by an suggestion by IBM"; it should be "a suggestion". The caption also has no final period.
- p. 47 Figure 9 caption
inspired by an suggestion by IBM
Judge's note: The p.47 crop confirms 'an suggestion'. The caption ends with 'than memory bandwidth' and no period.
Second check: none (judge only)
Report a problem with this finding
Typo Wrong word in the Conclusion: "an open, exciting, and urgent research agenda to flush out the concepts" should be "flesh out the concepts".
- p. 47 Conclusion, final paragraph
research agenda to flush out the concepts
Judge's note: The p.47 crop confirms 'flush out'. From the context the meaning is to develop the concepts, which is 'flesh out'.
Second check: none (judge only)
Report a problem with this finding
Typo Misspellings in the reference list: "Unrealiable", "C. Hewiit", "D. Cllahan", "Implemention", "Phsyics", "Eigth", "Auto-Partiontioning", "Internation Symposium", "Architectual Support", "Morgan Kauffman", "V. Luchango".
- p. 49 References, Borkar 2005
Unrealiable Components
- p. 51 References, Hewitt et al 1973
C. Hewiit
- p. 48 References, Alverson et al 1990
D. Cllahan
- p. 50 References, Dean and Ghemawat 2004
Design and Implemention
Judge's note: I checked each word on the crops for pp.48–55, and each is printed as quoted. The paper itself shows that 'Cllahan' is wrong: the next entry, Alverson et al 1999, prints 'Callahan'. 'Luchango' (for Luchangco) and 'Kauffman' (for Kaufmann) rest on outside knowledge of the correct spellings. The others are plainly misspelt English words.
Second check: none (judge only)
Report a problem with this finding
Typo The author list of [Wawrzynek et al 2006] includes "J.C. Joe"; other entries in the same reference list name this RAMP co-author as "J.C. Hoe".
- p. 55 References, Wawrzynek et al 2006
C. Kozyrakis, J.C. Joe, D. Chiou
Judge's note: The p.55 crop shows 'J.C. Joe'. In the text layer, [Arvind et al 2005] and [Chung et al 2006] both have 'J.C. Hoe', so the paper itself shows the misspelling.
Second check: none (judge only)